Not long ago, I pulled my smartphone from my pocket, unlocked the screen, and selected the icon of a nutrition analysis app. The app’s splash screen flashed into view, greeting me with a clear logo set against a crisp backdrop, and I waited patiently for the app to finish loading. But instead of transitioning to the app proper, to the tables of calories and proteins and carbohydrates, my phone paused on the splash screen. The few seconds of loading turned to tens of seconds, then to the better part of a minute, until finally the bright screen winked into a lifeless, unlit black. In only moments, my routine action—opening the app—morphed into something very different: a crash, a breakdown, a failure. Of course, my experience with a malfunctioning smartphone is, as an isolated incident, hardly noteworthy or even unusual. As anyone who has used a digital device or platform for any length of time can attest, digital media break, and they do so regularly. Indeed, failure is so much a part of the basic experience of digital media that it is only the most egregious breakdowns that draw users’ attention beyond a momentary burst of frustration. My smartphone’s failure was as routine and inconspicuous as the action that triggered it.
Given this routine disregard for failure and breakdown in users’ everyday experience, it is all too easy for scholars of digital media to follow a similar path by presuming an idealized object of study, one that functions perfectly, as intended, without errors, glitches, or bugs. In this way, failure in digital media can be ignored as the exception to the rule, and therefore external to the immediate bounds of inquiry. However, a recent wave of media studies scholarship has attempted to avoid such idealization by accounting for the myriad of ways that digital media ‘go awry.’ For example, instead of focusing on ‘functioning’ computer networks, Jussi Parikka’s Digital Contagions attends to the disruptions of computer viruses.1 In place of well-meaning email messages, Finn Brunton turns to the excesses and obscenities of spam.2 Peter Krapp’s Noise Channels centers on the disturbance of communication rather than its facilitation,3 and two edited collections—The Spam Book and Error—address a wide variety of digital breakdowns.4
Aesthetic considerations of digital failure have drawn an even larger share of scholarship, with several studies productively exploring the aesthetic role of the ‘glitch’.5 Furthermore, recent studies of the repair of technologies and infrastructures take the ubiquity of breakdown as their starting premise.6
Yet it would be a mistake to depict this flurry of scholarship on digital failure as conceptually homogenous. Even though these scholars are united by a desire to analyze failure, there are significant divergences over how such failures should be identified and understood. Put differently, while it is clear that studies of spam, contagions, noise, errors, bugs, and glitches all share a common orientation towards often overlooked phenomena, the relationship between these varied terms is far from self-evident. Indeed, at first glance this conceptual diversity seems to be a substantial stumbling block for further inquiry—an indication of some fundamental incommensurability between different avenues of scholarship. But diversity is not necessarily a limitation, and from another perspective it appears less as an obstruction and more as an opportunity. Simply put, this diversity begs the question, ‘What differentiates one concept from another?’
The central premise of the present essay is that this question can illuminate not simply the conceptual differences between various scholarly camps, but also something about the nature of digital failure itself, something about this nebulous thing variously labeled ‘error,’ ‘glitch,’ and so on. In what follows, I intervene in the burgeoning literature on failure by arguing that the differences highlighted by two distinct conceptions of digital failure—failure as anomaly and failure as error—bring into focus a hitherto unexplored rhetorical dimension to failure in digital media. More specifically, I develop this rhetorical conception of digital failure in two ways: First, I argue that failure is not simply a natural or logical extension of the structure of media objects, but rather is itself a rhetorical construction produced by discourses typically considered to be purely technical—for example, error messages. Second, I argue that these types of discourses should be understood as attempts to rhetorically enforce a kind of digital decorum—that is, distinctions between the ‘proper’ and ‘improper’ in digital media.
To advance these arguments, I first review the introductions of two edited collections on the topic of digital failure—Jussi Parikka and Tony D. Sampson’s The Spam Book, and Mark Nunes’ Error—and show how the former conceptualizes digital failure as anomaly, while the latter understands failure through the lens of error. The key conceptual difference between anomaly and error, I argue, lies in their temporal orientation: Anomalies are established by situating a given function with regard to a historical precedent, whereas errors are retroactively designated as such. Turning to the question of how anomalies and errors come into being leads me to argue that such failures are not simply self-evident states of technical malfunction, but are rather the products of two distinct rhetorical strategies—namely, preemptive and retrospective appeals. To the latter strategy I pay special attention, and I argue that such retrospective rhetoric works by supplementing other rhetorics already active within a given digital device or platform—that is, by encouraging a ‘proper’ interpretation of these other rhetorical appeals—and thereby contributes to a broader rhetorical ecology of decorum in digital media. Finally, I close by considering the implications of this supplementary form of persuasion for future studies of digital failure and for rhetorical analyses of digital media more broadly.
In their editorial introduction to The Spam Book, Jussi Parikka and Tony D. Sampson begin by noting that the phenomena addressed by their collection—spam, computer viruses, Internet pornography, and so on—clearly evoke negative associations.7 Given these associations, Parikka and Sampson argue, it is tempting to view these phenomena as a kind of ‘digital pollution’—to think of spam, viruses, and porn as ‘a major downside (or setback) to a communication revolution that promised to be a noiseless and friction-free Road Ahead.’8 But instead of turning to the ecological metaphor of pollution, Parikka and Sampson hinge their essay on another term: ‘Against the prescribed and often idealized goals of the visionaries of digital capitalism,’ spam and similar phenomena ‘appear to us as anomalies.’9
Why choose anomaly over pollution? On one hand, it is clear that anomaly is considerably less metaphorically charged than pollution; the latter contains a degree of moral judgment not as obvious in the former. But is anomaly really that neutral of a term? The conventional understanding of anomaly as deviation, while certainly not as salient in its biases as pollution, still implies some distance from a particular norm. At best, anomaly speaks to typicality, to statistical considerations. But if this is the case, anomaly seems a strange choice to describe spam and porn, given that these phenomena are hardly foreign to digital culture.10 Yet Parikka and Sampson are well aware of this apparent paradox:
If we constrain ourselves to the dictionary definition of the anomalous, as the unequal, uncomfortable, dissimilar, and incongruous, in other words, something that deviates from the rule and demonstrates irregular and abnormal behavior or patterns, then arguably our question [i.e., how are objects like spam anomalous?] becomes problematized by everyday experiences of network culture.11
Parikka and Sampson’s anomalies, then, are not so anomalous in the conventional sense. And this, in a roundabout way, turns out to be precisely the point with which the pair grapples: How should one analyze phenomena that seem to be out of place in the discursive or ideological construction of network culture (i.e., aberrations with respect to the ‘idealized goals of the visionaries of digital capitalism’) yet are plainly not out of place in its everyday functioning?
Parikka and Sampson’s solution to this problem involves a move away from what they term ‘representational analysis’—that is, analysis of discourses about technologies. ‘Our aim as editors,’ they write, ‘has been to steer clear of the linguistic categorizations founded on resemblances, identities, oppositions, and metaphorical analogies.’12 Instead, they are
concerned with affect and ethology: how various assemblages of bodies (whether technological, biological, political or representational) are composed in interaction with each other and how they are defined, not by forms and functions, but by their capabilities or causal capacities.13
In other words, they propose to shift the line of inquiry from ‘What do these anomalies mean?’ to ‘What do these anomalies do?’ In this way, Parikka and Sampson follow Paul Virilio’s proposal to reverse the conventional view of accident as contingent and substance as necessary: Digital anomalies are necessary components of much larger assemblages, and any analysis of anomalies should treat them as such.14 Spam, for example, should not be viewed as some aberration within the communication system of email, but rather as part and parcel of that system, an element internal to it.
The ultimate payoff of this approach is not so much a robust theorizing of the anomalous in particular—that is, defining specific features unique to the anomaly—but rather a kind of democratization, an ontological ‘flattening’ or equalization of digital objects.15 For Parikka and Sampson, anomalies such as spam stand out as distinct not because they are inherently anomalous, but because we treat them as such. The solution, then, is to bracket the assumption that anomalies are somehow unequal or dissimilar. Thus, Parikka and Sampson’s choice of anomaly over pollution speaks less to a description of anomalous objects and more to a theoretical orientation that resists the urge to identify anomaly with exception. From Parikka and Sampson’s perspective, the way to evade ‘the binary impasse of the normal and the abnormal’ is to treat all digital phenomena—including failures and breakdowns—as equally legitimate participants in network culture.16
Mark Nunes’ introduction to Error parallels Parikka and Sampson’s essay in several respects. Like Parikka and Sampson, Nunes begins his introduction by considering ideology. ‘The network society,’ Nunes claims,
is governed by what Jean-Francois Lyotard defined more than thirty years ago as a ‘logic of maximum performance’: a cybernetic ideology driven by dreams of an error-free world of 100 percent efficiency, accuracy, and predictability.17
As it is for Parikka and Sampson, the problem for Nunes is that this ‘cybernetic ideology’ does not fully account for the failures and breakdowns of experience. And like Parikka and Sampson, Nunes takes issue with the lack of analysis dedicated to moments of digital breakdown, and sees the potential for such analysis to explicate workings of Lyotard’s error-free ‘logic of maximum performance’ in productive ways. Focusing on digital breakdowns, Nunes argues, ‘reveals not only a system’s failure, but also its operational logic.’18
But Nunes parts company from Parikka and Sampson on the particular term he uses to encapsulate digital failure. Rather than turning to anomaly, Nunes chooses error as his conceptual focus. Like Parikka and Sampson’s choice of anomaly over pollution, Nunes’ choice of error is initially difficult to understand. Given that Error appeared two years after the publication of The Spam Book, and that Nunes is plainly aware of Parikka and Sampson’s work—he cites their introduction to The Spam Book—why choose error over anomaly? Nunes’ preference for error over anomaly must be understood with respect to error’s unique etymological connection to wandering. Error, unlike anomaly, evokes ‘a going astray, a wandering from intended destinations.’19 And while anomaly’s first association is with atypicality, error—in the sense of straying—implies no degree of infrequency. Error does, however, suggest aspects of intention and purpose absent in anomaly. And this unintentional dimension of error, according to Nunes, opens a space of possibility vis-à-vis the aforementioned ‘logic of maximum performance’:
While often cast as a passive, yet pernicious deviation from intended results, error can also signal a potential for a strategy of misdirection, one that invokes a logic of control to create an opening for variance, play, and unintended outcomes. Error, as errant heading, suggests ways in which failure, glitch, and miscommunication provide creative openings and lines of flight that allow for a reconceptualization of what can (or cannot) be realized within existing social and cultural practices.20
The conceptual support for these error-inspired ‘creative openings’ comes from Nunes’ theorization of error via Claude Shannon’s information theory rather than Norbert Wiener’s cybernetics. In brief, Nunes locates the key difference between Shannon and Wiener in the ways that each relates error to larger systems of communication and control. Error, in Wiener’s cybernetic view, constitutes ‘a measure of the gap between “actual performance” and “expected performance”’; it is conceptualized as an actuality deviating from some ideal.21 In this way, error is integrated into systemic control though mechanisms of feedback, and ‘serves its purpose as a corrective—what keeps purpose on purpose and tasks on goal.’22 As Nunes points out, this view of error resonates strongly with functionalist theories of society: As in Émile Durkheim’s ‘society of saints,’ it is the very transgression of limits that establishes the norm.23 Thus, from the cybernetic perspective, errors exist not as phenomena external to the system, but rather as an integral part of its functioning.
It is worth noting that there are clear resonances between this cybernetic view of error and Parikka and Sampson’s formulation of failure as anomaly: Just as error is a part of the cybernetic system, and not external to it, anomalies are internal elements of network culture. But Nunes does not settle with this cybernetic understanding of error, and at this point, it is possible to discern the root of his divergence from Parikka and Sampson. Cybernetics’ self-integration of error ‘as captured, predictable deviation [that] serves order through feedback and systemic control’ effectively limits the emancipatory potential that Nunes attempts to locate within moments of digital failure.24 Thus, if error is to truly offer ‘creative openings,’ Nunes must turn to a different, non-cybernetic conception of error.
To move beyond cybernetics and its functionalist treatment of error, Nunes roots his own understanding of error in the work of Claude Shannon. As Nunes points out, Wiener’s cybernetics remains tied to a view of entropy that is opposed to information: The greater the amount of information (i.e., feedback), the greater the degree of certainty (i.e., error correction). Error in this sense is a form of information that reduces entropy, insofar as it participates in mechanisms of feedback. But Shannon’s equation of information with entropy—what N. Katherine Hayles has distinctively called ‘Shannon’s choice’25—inaugurates a radically different understanding of error. In Shannon’s view, error names not a deviation from an ideal, but rather the proliferation of potential inherent in information itself. A simple binary code, for example, contains the fewest possible interpretations (e.g., zero or one), and therefore the lowest level of entropy. Any increase in the code’s complexity (i.e., an increase in information) increases possible interpretations, and thus introduces a greater potential for message degradation. This conception of error as tied to potential returns to Nunes’ initial stress of error’s etymology: A ‘wandering from intention’ is an exploration of the potential that inhabits a given system.
Admittedly, this view of error as potential may seem to be just another means of addressing the same basic concept outlined by Parikka and Sampson: Both error and anomaly are concerned with failure not as a violation of some norm but as a phenomenon inherent to network culture—either as a portion of an assemblage, or as a potentiality of information. Yet this conceptual similarity belies a key temporal dimension that sets error apart from anomaly. Anomaly, as articulated by Parikka and Sampson, moves forward with respect to the past: The judgment of anomaly, whether in terms of atypicality or in Parikka and Sampson’s broader sense, is informed by a history—a comparison of one digital breakdown to a preexisting series. To return to this essay’s opening anecdote about smartphone failure: What makes the app’s failure to load so unremarkable and inconspicuous is its situation within a history of previous failures. In this way, an anomaly is an anomaly from the very moment of its appearance; as soon as the anomaly appears, I know what it is with respect to my past experience. The past, so to speak, produces the present.
Error may function according to a similar temporal logic—and it is this logical similarity that makes Nunes seem so close to Parikka and Sampson—but it also includes the potential for a very different temporal configuration. This is best demonstrated by way of example. Imagine that I want to send a text message to arrange lunch with a friend: ‘I’m very hungry. Let’s meet at noon.’ But also imagine that I am a bit clumsy, and instead of carefully selecting the keys necessary to spell ‘hungry,’ my fingers smudge a combination of keys less precise: s-n-g-r-y. My phone strains its algorithms to make the best of this jumbled input, and automatically corrects this to ‘angry,’ producing ‘I’m very angry. Let’s meet at noon.’ I send the message, my friend receives it, and understands it perfectly—all while missing what I intended to say.
In this case, error should not be located in any disruption of communication. There is no external entity interfering with the transmission of my message; the sending and decoding of the message happens flawlessly. Even my encoding—my clumsy typing, filtered through my phone’s algorithms—is not a disruption of communication. The ‘intervention’ of autocorrect is not qualitatively different than a more ‘direct’ translation of my typing into text. It is considerably more complex, to be sure, insofar as it takes into account a much broader range of input (e.g., the history of my typing), but it is still an essentially mechanistic process, no different in kind from the operation of a mechanical typewriter. Together, my phone’s keys and its algorithms produce exactly the message that I type, even if it is not what I meant to type. Thus, the error of this example exists not as a disruption of communication, but rather as a disjunction between the intended message and the actual message.
But when does this disjunction actually come into being? That is to say, when is the separation of intention and actuality first detectable? Not when my friend receives my message, since it is very easy to imagine him taking ‘I’m very angry’ as my intended meaning. Not at the moment that I send the text. Perhaps when my fingers hit the wrong buttons? Yet it is not fair to refer to them as the ‘wrong’ buttons before establishing error’s existence, since that merely begs the question. It seems the only possible answer is that error is brought into existence by its recognition.26 All of this happens retroactively; it is only when error is seen as such that it will have existed. In the absence of such retroactive recognition, an error is simply another potential—and equally valid—interpretation of a message.
The construction of failure
This divergence between the temporal logics of anomaly and error highlights an important facet of digital failure: The designation of failure as such is not self-evident, and instead requires a form of work. At first, this claim may seem counterintuitive, especially when held against Parikka and Sampson’s discussion of anomaly. Is not the whole point of Parikka and Sampson’s choice of anomaly over pollution that digital disruptions are intuitively ‘bad’? The turn to topologies and assemblages seems to assume that the ‘bad’ view of anomalies is the default, the state sans labor.27 Yet even if Parikka and Sampson are right in considering anomalies apart from their negative connotations, this does not mean that these negative connotations are inherent to anomalies. Indeed, the key contribution of Nunes’ theory of error seems to be the way in which it points to the possibility of historicizing anomaly, of providing an account of why Parikka and Sampson’s refiguring of anomaly might be necessary in the first place. With this in mind, it is clear that digital failure must be understood not only in terms of Parikka and Sampson’s anomaly—although this is certainly a useful approach—but also as something that is produced, the product of some labor.
So what is the nature of this labor? To answer this question, it is useful to return to the previously mentioned text message example. If my message of ‘I’m very angry. Let’s meet at noon.’ only becomes an error when it is recognized as such, how does this recognition happen? As the sender of this message, I recognize the error as soon as I receive my own message—that is, as soon as I read my own text and notice its divergence from my intent. My friend, however, can only recognize the error with certainty if I somehow communicate a supplementary message that designates the first as erroneous. Furthermore, my friend may not necessarily accept my explanation of error; he may, for example, see ‘angry’ as indicative of some Freudian mechanism rather than of my unwieldy thumbs. Regardless, it is clear that on some level I must persuade my friend that an error has occurred. The identification of error is therefore a deeply rhetorical process.28
In the context of digital media, the identification of error usually involves at least one of two rhetorical strategies. The first and perhaps more familiar approach is to preemptively establish certain functions as failures, by setting user expectations in advance of the failure actually occurring. This form of rhetoric can most often be seen in sources external to the device or platform under consideration—for example, in user manuals, advertisements, the advice and instruction of other users, as well as in generalized patterns of design that may transcend specific media or platforms (e.g., the adaptation of the physical top of a desk to the digital ‘desktop’). This particular strategy is not, of course, unique to digital media, and there are a number of similar examples to be found in other contexts. The rhetoric of a videogame user manual, for example, attempts to establish user expectations relative to the game, and in this way is not substantially different from the manual of any other, non-digital artifact.
The second strategy, by contrast, is more uniquely situated within digital media. While the first rhetorical strategy involves discourse external to the device or platform in question, the second strategy involves the capacity of digital objects to advance arguments about themselves. This is most apparent in the case of the error message, in which the digital object gives an account of its own functioning. When my web browser displays a HTTP 404 ‘Not Found’ error message, this is quite explicitly a claim that some aspect of its functioning is disrupted—namely, that the server was unable to locate what the client requested. Furthermore, this error message refers not to something that is to come, but to something that has already occurred, and thus utilizes a temporal logic that differs substantially from the rhetoric of the user manual and other objects that employ preemptive strategies. The temporal perspective at work in the error message is not preemptive, but retrospective. The error message appears only after the fact, as a notice that something has already gone wrong.
Notably, error messages work not only by making explicit statements (e.g., displaying a particular error code), but also by invoking particular visual aesthetics. Perhaps the best example of this dynamic is the famous ‘Blue Screen of Death’ message utilized by all Microsoft Windows operating systems since Windows 3.1.29 The Blue Screen of Death displays, in simple white text on a bold blue background, a message addressed to the user, usually some variant of ‘A problem has been detected and Windows has been shut down to prevent damage to your computer,’ followed by technical information. And while the Blue Screen quite obviously serves a diagnostic function in a technical sense—various error codes are displayed to pinpoint the problem, for example—the visual aesthetic of the error code (e.g., the white and blue color scheme) is itself so recognizable as an indicator of error that a simple Google search of ‘blue screen of death’ reveals countless adaptations and parodies: One is refashioned as a Budweiser advertisement, another humorously rephrases the technical jargon in excessively dramatic terms, and yet another distills the message to its bare aesthetic by replacing the text with a string of As (e.g., ‘AAAAAAAA’).
Still, it would be a mistake to downplay the diagnostic dimensions of the error message. Such messages often provide valuable information regarding the nature of an error, and in most cases any argument for the existence of error as an error is more or less redundant; the user manuals, advertisements, and other preemptive rhetorics have already done their work, and the user who encounters a failure is already well convinced that something has gone wrong.30 Yet even if the identification of error (i.e., retrospective rhetoric) proves, in most cases, to be superfluous or redundant, it is not so in all cases. There is an ambiguity here, an excess of potential—the same slippage that Nunes identifies with error. The qualification of ‘in most cases’ points to the fact that some failures are contested. Sometimes, despite the efforts of preemptive rhetorics and retrospective rhetorics, the user is unconvinced that any sort of error has occurred.
To fully understand these contested moments, it is helpful to take a short detour through a related branch of rhetorical theory. Ian Bogost has argued that the very structure of digital media objects can itself be considered a form of persuasion.31 Such objects utilize what Bogost calls ‘procedural rhetoric’—that is, persuasion via the representation of rule-based systems, systems that afford some actions on the part of a user, and constrain others. Though Bogost mainly uses procedural rhetoric to explicate videogames, the concept can very easily be generalized beyond the genre of the game: Just like videogames, operating systems and word processors and smartphones—in short, all digital media devices and platforms—enact certain rhetorics through their structuring of user experiences and avenues of action. But procedural rhetorics, just like their discursive counterparts, do not always function as their authors intend. Just as any speech raises the possibility of unintended interpretation—Mitt Romney’s ‘binders full of women’ remark stands out as a recent and memorable example of this in a more traditional setting32—so do procedural rhetorics offer opportunities for subversion.
This is perhaps best illustrated by way of Bogost’s preferred genre, the videogame. In the hit videogame series Tomb Raider, for example, players control adventurer Lara Croft as she runs, jumps, swims, and shoots her way through a number of varied environments—caves, ancient temples, and so on.33 Importantly, even though Lara is capable of extraordinary physical feats—she can, for example, leap across a large chasm only to catch a tiny ledge by her fingertips—the narrative of the games ultimately characterizes her as an ordinary human; she has no magical powers, and all of her abilities are at least somewhat plausibly rooted in her characterization as an (extraordinarily athletic) adventurer.
However, in the early entries of the Tomb Raider series, players discovered a ‘bug’ in the game engine—a bug that affords Lara abilities not supported by the Tomb Raider narrative.34 If players maneuver Lara to stand near the base of certain walls, and then cause Lara to jump, the Tomb Raider game engine may ‘mistakenly’ teleport Lara to the top of the particular wall—a feat not possible by ‘normal’ means within the game. Yet it is worth noting that, in one sense, such teleportation did not actually break any of the game’s ‘rules.’ If, following Parikka and Sampson, this bug is considered as an anomaly, rather than as some ‘flaw,’ then the gameplay possibilities afford by the bug (e.g., accessing normally restricted areas) should be considered as equal to the gameplay possibilities explicitly supported by the game’s narrative (e.g., proceeding through the space of the game along a particular path). What this particular bug does, then, is point to a disjunction between the procedural rhetoric of the game—that is, the ‘rules’ governing gameplay—and the discursive rhetoric of the game—that is, its narrative. Players who teleport Lara to the top of walls play by the game’s rules, but not by the game’s narrative.
Importantly, players of the early Tomb Raider games do not necessarily view these bugs as moments of failure within the game, even though they quite clearly depart from the game’s narrative trajectory. As one Tomb Raider fan website states:
NOT ALL BUGS ARE BAD. Like most computer/video games, the Tomb Raider series has its share of glitches. Some are mildly annoying, some are deadly—to Lara, that is. […] But there are a few that can actually enhance your gaming experience.35
Clearly, the designers of the Tomb Raider series had no intention of introducing a bug that allowed Lara to teleport to the top of buildings—and from that perspective, the bug is a failure. But at least some players contest this interpretation, and take a quite different view of the bug: From this view, not all bugs are failures.36
Although there are no error messages active in the Tomb Raider example, the fact that the meaning of bugs may be contested helps to situate the rhetorical role of error messages in other media objects. If, following Parikka and Sampson, moments of failure are not assumed to be inherently ‘bad,’ and are instead read as unanticipated functions that subvert the normal, procedural rhetoric of the device or platform, then error messages become legible as a kind of supplementary rhetoric—a discursive ‘add-on’ meant to convince users that a particular form of functioning was not supposed to happen. In this way, error messages, though undoubtedly designed with the intent of communicating useful technical information, actually end up performing a secondary role: They rhetorically enforce particular divisions between the ‘proper’ and ‘improper’ functioning of a given digital object, regardless of the actual functioning of the object and the user’s interpretation of that functioning. In other words, error messages communicate not only technical details, but also an entire politics; they are a means of managing the potential ‘unruliness’ of users, the ‘creative openings’ theorized by Nunes.
Yet this political dimension to error messages seems, at first glance, in conflict with this essay’s opening observation: Digital media fail routinely, and these failures are generally not considered to be noteworthy or noticeable by users. So how it is possible to reconcile the politics and explicit persuasion of the error message with the end result of error’s effective invisibility? To answer this question, it is useful to turn to another rhetorical concept: to prepon. Variously translated as decorum or propriety, to prepon is the sense of what is stylistically appropriate for a given combination of speaker, subject, audience, and context.37 Given that no two configurations of these elements are absolutely identical, decorum necessarily exceeds that which can be stated in a prescriptive form; the variability of context cannot be captured in a single rule or statement. As Richard Lanham notes in The Electronic Word, decorum is thus ‘that good taste which cannot be taught but yet somehow must be learned, since it lurks at the center of everything.’38 Lanham also notes that this apparent paradox draws attention to decorum’s relationship to self-consciousness: Precisely because it is felt intuitively and not stated explicitly, decorum—as a ‘permitted range of dynamic variation’ in the relationships between speaker, subject, audience, and context—is what ‘we must interiorize but not bring to self-consciousness.’39 In other words, decorum may be said to be operative in a given communicative instance precisely to the extent that participants in that instance internalize it—that is, to the extent that it eludes their conscious attention.
With these two elements in mind—the irreducibility of decorum to a single prescription, and its relationship to self-consciousness—it is possible to return to the political stakes of error messages. If such messages are understood not only as attempts to label particular functions as errors, but also as support for a kind of digital decorum, then it is possible to identify two key features of the rhetorical work of error messages: First, as contributions to a particular sense of decorum, error messages can be understood not only as individual, standalone statements about a particular moment of (mal)functioning, but also as elements of a larger rhetorical ecology, an ecology focused on the divide between the proper and the improper. In this way, the rhetoric of error messages may be understood as working in concert with user manuals, advertisements, generic constraints, procedural structures, and the like. Indeed, it is this broader rhetorical ecology of decorum that can make failures visible even in the absence of explicit error messages (e.g., the Tomb Raider teleportation bug, as a bug, emerges out of a particular decorum of gameplay that is reinforced by the game’s narrative, genre, manual, etc.). Second, as contributions to this larger rhetorical ecology of decorum, and given that decorum is necessarily internalized, error messages can be understood as constructing not only the ‘proper’ functioning of digital objects, but also ‘proper’ users as a part of this functioning.
At the most basic level, this construction of the ‘proper’ user can be seen in the particular affordances and information provided to users by error messages. Variations in either—for example, the ability to submit an online error report, or the inclusion of detailed error codes—situate the user as more or less invested in the maintenance and operation of a given digital object. In other words, every error message is premised on a particular understanding of the user’s role relative to the functioning of the object.
And in one sense, such varied conceptions of the user are hardly surprising. At the level of development and production, for example, different design philosophies conceptualize different ideal users,40 and the debates inspired by the free software movement are explicitly centered on the ways in which digital objects structure the role of users as more or less passive with regard to the overall functioning of the those objects. But beyond this more explicitly ideological level, error messages also shape users in more subtle ways by placing users in a passive position relative to digital failure specifically. The retrospective nature of the rhetoric of error messages disrupts the connection between user agency and the error itself: The error message not only attempts to persuade the user that an error has already occurred, but also situates errors as things that happen to users, as opposed to being caused by users.
Nevertheless, this construction of users as passive in relation to errors does not simply restrict the agency of users relative to the object. To the contrary, it sustains both an affective relief to the frustration of encountering digital failure, and a range of social practices centered on the use of a given digital object. By locating the failure in the device itself rather than in the actions of the user, the error message absolves the user of responsibility for the error, and in turn, this enables users to continue patterns of behavior that may in fact be more likely to trigger further moments of failure and disruption. This dynamic can even be seen in the absence of automatically generated error messages: When I text my friend, for example, the very existence of autocorrect as an entity which may bear responsibility for my clumsy mistakes contributes to my ability to be less concerned about the construction of my messages—to text quickly while multitasking, and so on. In this case, errors are constructed not simply as my fault but as the fault of the device itself (i.e., autocorrect). As such, errors enable particular social configurations: I am able to participate in practices—such as rapid, inattentive texting—that treat errors as simply ‘bound to happen.’
Thus, the rhetoric of error exhibits not only a persuasive function (i.e., users must be convinced that an error did in fact occur) but also, insofar as it enacts a sense of decorum, a constitutive function: Users must be constructed as an audience, as ‘proper’ users and as participants in particular sets of practices centered on the ‘proper’ functioning of a given digital object. Of course, individual users bring their own idiosyncrasies to interactions with digital objects, and there is no guarantee that one element in a rhetorical ecology of decorum (e.g., an error message) will work in perfect concert with the rest, but ultimately to be a user is to be subject to a particular sense of decorum, to be able to see through a particular rhetorical ecology to intuitively grasp the appropriate functioning of user and object. In this way, it is possible to reconcile the persuasive rhetoric of the error message with error’s invisibility in the everyday experience of users: The very fact that the error is invisible is a testament to the efficacy of this rhetoric, of the extent to which users have internalized a particular decorum, and in doing so have become constituted as ‘proper’ users.
Error, rhetoric, and media
Rhetorical studies of media technologies generally fall into one of three camps. The first considers rhetoric about technologies—that is, the ways in which technological artifacts are constructed through discourses and interpretations.41 The second considers the ways in which a given technology affects or alters particular modes of persuasion.42 The third considers how technologies themselves may be persuasive in a non-discursive manner—procedural rhetoric, for example, fits well within this vein. Yet close attention to the rhetoric of error in digital media challenges these clear-cut distinctions. Discourses about technologies are not isolated from procedures within technologies; as the rhetoric of error messages demonstrates, the procedural rhetorics embedded within media objects may be supplemented by other, more explicitly discursive rhetorics, and these may work together in a larger rhetorical ecology. Thus, moments of failure in digital media have much to offer scholars concerned with the rhetorical work of media objects.
However, the reverse is also true. Scholars who focus on moments when digital media ‘break down’ stand to benefit from considering the ways in which these breakdowns come into being, of the ways in which anomalies come to be intuitively ‘bad’, and the ways that users come to understand certain functions as errors. The gesture—performed in different ways by The Spam Book and by Error—of bracketing assumptions about digital failures is a valuable move for the larger domain of media scholarship, but media theorists should not forget to examine the conditions that make such a gesture necessary in the first place—that is, the conditions that produce the decorous divide between functioning and malfunctioning. And it is here, in the analysis of this intuitive and inconspicuous divide, that rhetorical scholarship can be especially helpful to media theory. As Richard Lanham writes, decorum is that which we must ‘not bring to self-consciousness,’
and yet anyone studying language must do this, must break down the compelling urge to see through our means of seeing to the ‘reality’ established by that seeing. Fabrication of the ‘decorous,’ unselfconscious Western reality, stylistic or social, is done through a trick, a series of tricks, just like perception itself, and we want to know how the trick is done.43
In turn, the task of rhetorical analysis is to ‘[make] us look AT the verbal surface rather than THROUGH it to the “reality” our decorous trickery has created.’44 Applied to cases of digital failure, such a method asks us to look at the means by which failures are produced, not simply through those means to take failures as given.
As the burgeoning literature on digital failure has demonstrated, to presume an idealized media object, one that functions ‘properly’—that is, without failures and errors—ignores the basic experience of digital media as objects that quite regularly break. But this presumption also overlooks the extent to which digital media objects are deeply rhetorical in nature, and exist as part of broader rhetorical ecologies. To focus only on discourse external to objects—what Parikka and Sampson term ‘representational analysis’—ignores the ways in which objects advance their own rhetorics, both in procedure and in their own discursive self-commentary. Likewise, any analysis limited only to procedural rhetoric will necessarily miss the ways in which such rhetoric may be supplemented at a discursive level by rhetorics external and internal to the media object. In short, any comprehensive study of digital media objects must account for the ways in which those objects fail, and any comprehensive account of those failures must attend to the ways in which such failures are constructed by a diverse ecology of rhetorical appeals.
‘Autocorrect.’ Know Your Meme. http://knowyourmeme.com/memes/autocorrect.
Anthony, Sebastion. ‘Just How Big Are Porn Sites?’ ExtremeTech, April 4, 2012. http://www.extremetech.com/computing/123929-just-how-big-are-porn-sites.
‘Blue Screen Data.’ MSDN Library, April 9, 2013. http://msdn.microsoft.com/en-us/library/windows/hardware/ff538869(v=vs.85).aspx.
Bogost, Ian. Persuasive Games: The Expressive Power of Videogames. Cambridge, MA: MIT Press, 2007.
Brunton, Finn. Spam: A Shadow History of the Internet. Cambridge, MA: MIT Press, 2013.
Cardona, Maria. ‘Romney’s Empty “Binders Full of Women.”‘ CNN.com, October 18, 2012. http://www.cnn.com/2012/10/17/opinion/cardona-binders-women/index.html.
Damn You Auto Correct! http://www.damnyouautocorrect.com.
Durkheim, Émile. The Rules of the Sociological Method. Edited by Steven Lukes. Translated by W.D. Halls. New York: Free Press, 1982.
Gates, Bill. The Road Ahead. London: Penguin, 1996.
Graham, Stephen, and Nigel Thrift. ‘Out of Order: Understanding Repair and Maintenance.’ Theory, Culture & Society 24, no. 3 (2007): 1-25. doi:10.1177/0263276407075954.
Hayles, N. Katherine. Chaos Bound: Orderly Disorder in Contemporary Literature and Science. Ithaca: Cornell University Press, 1990.
Jackson, Steven J. ‘Rethinking Repair.’ In Media Technologies: Essays on Communication, Materiality, and Society, edited by Tarleton Gillespie, Pablo J. Boczkowski, and Kirsten A. Foot, 221-240. Cambridge, MA: MIT Press, 2014.
Kelly, Caleb. Cracked Media: The Sound of Malfunction. Cambridge, MA: MIT Press, 2009.
Kline, Ronald and Trevor Pinch. ‘Users as Agents of Technological Change: The Social Construction of the Automobile in the Rural United States.’ Technology and Culture 37, no. 4 (1996): 763-95. doi:10.2307/3107097
Krapp, Peter. Noise Channels: Glitch and Error in Digital Culture. Minneapolis: University of Minnesota Press, 2011.
Lanham, Richard A. The Electronic Word: Democracy, Technology, and the
Arts. Chicago: University of Chicago Press, 1993.
Massanari, Adrienne L. ‘Designing for Imaginary Friends: Information Architecture, Personas and the Politics of User-centered Design.’ New Media & Society 12, no. 3 (2010): 401–16. doi:10.1177/1461444809346722.
McKenna, Stephen. Adam Smith: The Rhetoric of Propriety. Albany: SUNY Press, 2006.
Montfort, Nick, and Ian Bogost. Racing the Beam: The Atari Video Computer System. Cambridge, MA: MIT Press, 2009.
Nunes, Mark. ‘Error, Noise, and Potential: The Outside of Purpose.’ In Error: Glitch, Noise, and Jam in New Media Cultures, edited by Mark Nunes, 3–23. New York: Continuum, 2011.
———. , ed. Error: Glitch, Noise, and Jam in New Media Cultures. New York: Continuum, 2011.
Parikka, Jussi. Digital Contagions: A Media Archaeology of Computer Viruses. New York: Peter Lang, 2007.
Parikka, Jussi, and Tony D. Sampson. ‘On Anomalous Objects of Digital Culture: An
Introduction.’ In The Spam Book: On Viruses, Porn, and Other Anomalies from the Dark Side of Digital Culture, edited by Jussi Parikka and Tony D. Sampson, 1–18. Cresskill, NJ: Hampton Press, 2009.
———. , eds. The Spam Book: On Viruses, Porn, and Other Anomalies from the Dark Side of Digital Culture. Cresskill, NJ: Hampton Press, Inc., 2009.
Stellalune. ‘Classic Tomb Raider Useful Bugs.’ Stella’s Tomb Raider Site. Accessed March 15, 2014. http://tombraiders.net/stella/goodbugs.html.
Tomb Raider. Derby, UK: Core Design, 1996.
Virilio, Paul. ‘The Primal Accident.’ In The Politics of Everyday Fear, edited by Brian Massumi, 211–18. Minneapolis: University of Minnesota Press, 1993.
Warnick, Barbara, and David S. Heinemen. Rhetoric Online: The Politics of New Media. 2nd ed. New York: Peter Lang, 2012.
- Jussi Parikka, Digital Contagions: A Media Archaeology of Computer Viruses (New York: Peter Lang, 2007). ↩
- Finn Brunton, Spam: A Shadow History of the Internet (Cambridge, MA: MIT Press, 2013). ↩
- Peter Krapp, Noise Channels: Glitch and Error in Digital Culture (Minneapolis: University of Minnesota Press, 2011). ↩
- Jussi Parikka and Tony D. Sampson, eds., The Spam Book: On Viruses, Porn, and Other Anomalies from the Dark Side of Digital Culture (Cresskill, NJ: Hampton Press, 2009); and Mark Nunes, ed., Error: Glitch, Noise, and Jam in New Media Cultures (New York: Continuum, 2011). ↩
- For example, see Caleb Kelly, Cracked Media: The Sound of Malfunction (Cambridge, MA: MIT Press, 2009). ↩
- See Stephen Graham and Nigel Thrift, ‘Out of Order: Understanding Repair and Maintenance,’ Theory, Culture & Society 24, no 3. (2007): 1-25, doi:10.1177/ 0263276407075954; and Steven J. Jackson, ‘Rethinking Repair,’ in Media Technologies: Essays on Communication, Materiality, and Society, ed. Tarleton Gillespie, Pablo J. Boczkowski, and Kirsten A. Foot (Cambridge, MA: MIT Press, 2014), 221-240. ↩
- Jussi Parikka and Tony D. Sampson, ‘On Anomalous Objects of Digital Culture: An Introduction,’ in The Spam Book: On Viruses, Porn, and Other Anomalies from the Dark Side of Digital Culture, ed. Jussi Parikka and Tony D. Sampson (Cresskill, NJ: Hampton Press, 2009), 3. ↩
- Ibid. The allusion is to Bill Gates, The Road Ahead (New York: Penguin, 1996). ↩
- Parikka and Sampson, ‘On Anomalous Objects of Digital Culture,’ 3 (italics original). ↩
- According to one widely cited estimate, porn alone comprises around thirty percent of all Internet traffic. See Sebastion Anthony, ‘Just How Big Are Porn Sites?,’ ExtremeTech, April 4, 2012, http://www.extremetech.com/computing/123929-just-how-big-are-porn-sites. ↩
- Parikka and Sampson, ‘On Anomalous Objects of Digital Culture,’ 3 (italics original). ↩
- Ibid., 4. ↩
- Ibid., 5 (italics original). ↩
- Paul Virilio, ‘The Primal Accident,’ in The Politics of Everyday Fear, ed. Brian Massumi (Minneapolis: University of Minnesota Press, 1993), 211-18. ↩
- This resonates with platform studies’ integration of digital media and object-oriented ontology. For example, see the afterword of Nick Montfort and Ian Bogost, Racing the Beam: The Atari Video Computer System (Cambridge, MA: MIT Press, 2009). ↩
- Parikka and Sampson, ‘On Anomalous Objects of Digital Culture,’ 4. ↩
- Mark Nunes, ‘Error, Noise, and Potential: The Outside of Purpose,’ in Error: Glitch, Noise, and Jam in New Media Cultures, ed. Mark Nunes (New York: Continuum, 2011), 3. ↩
- Ibid. ↩
- Ibid. ↩
- Ibid., 3-4 (italics added). ↩
- Ibid., 6. ↩
- Ibid., 7 (italics original). ↩
- See Émile Durkheim, The Rules of the Sociological Method, ed. Steven Lukes, trans. W.D. Halls (New York: Free Press, 1982), 85-107. ↩
- Nunes, ‘Error, Noise, and Potential,’ 12 (italics original). ↩
- See N. Katherine Hayles, Chaos Bound: Orderly Disorder in Contemporary Literature and Science (Ithaca: Cornell University Press, 1990). ↩
- This dynamic helps to illuminate a related phenomenon: Internet memes that focus on autocorrect errors. Websites like Damn You Auto Correct!, which document moments of autocorrect-enabled miscommunication, are enjoyable precisely because they exploit this retroactive recognition of error: We laugh when we recognize the error, but also when we see the recognition of that error by the original sender or the receiver of the message. See ‘Autocorrect,’ Know Your Meme, http://knowyourmeme.com/memes/autocorrect; and Damn You Auto Correct!, http://www.damnyouautocorrect.com. ↩
- Similarly, studies of repair take breakdown as the given upon which the labor of repair acts. As Graham and Thrift write, it is the ‘processes of maintenance and repair’ that serve as ‘the main means by which the constant decay of the world is held off.’ See Graham and Thrift, ‘Out of Order,’ 1. ↩
- Of course, in most everyday contexts, such rhetoric functions without much of a hitch. Freud is again a useful reference to illustrate this point: The psychoanalytic setting is unusual precisely because it disallows the everyday rhetoric of ‘I didn’t’ mean that!’ But the quotidian nature of such rhetoric does not mean that it is not worth examining. To the contrary, the ease with which such rhetoric is overlooked—if anything, an indicator of its efficacy—makes it all the more intriguing. ↩
- More formally, the Blue Screen of Death is known simply as a ‘blue screen,’ a ‘bug check screen,’ or a ‘stop screen.’ See ‘Blue Screen Data,’ MSDN Library, April 9, 2013, http://msdn.microsoft.com/en-us/library/windows/hardware/ff538869. ↩
- The preemptive role of the user manual becomes even clearer when the audience of the manual is taken into account. A user well-accustomed to a similar device or piece of software may not read the user manual at all, precisely because he or she already possesses a set of expectations regarding ‘failure’ and ‘success.’ By contrast, a user without such expectations is more likely to consult the manual. In this sense, to become an ‘experienced’ user is to become familiar not only with particular technologies, but also with particular expectations for the functioning of those technologies. ↩
- Ian Bogost, Persuasive Games: The Expressive Power of Videogames (Cambridge, MA: MIT Press, 2007). ↩
- Maria Cardona, ‘Romney’s Empty “Binders Full of Women,”’ CNN, October 18, 2012, http://www.cnn.com/2012/10/17/opinion/cardona-binders-women/index.html. ↩
- The first entry in the series is Tomb Raider (Derby, UK: Core Design, 1996). ↩
- For amore thorough description of the bug as it appears in various entries of the Tomb Raider series, see ‘Classic Tomb Raider Useful Bugs,’ Stella’s Tomb Raider Site, accessed March 15, 2014, http://tombraiders.net/stella/goodbug.html. ↩
- Ibid. ↩
- This offers a new reading of the old joke concerning bugs: ‘It’s not a bug; it’s an undocumented feature!’ Perhaps error messages are best understood as an inversion of this: ‘It’s not an undocumented feature; it’s a bug!’ ↩
- For a useful overview of the intellectual history of this concept in the rhetorical tradition, see Stephen McKenna, Adam Smith: The Rhetoric of Propriety (Albany: SUNY Press, 2006). ↩
- Richard A. Lanham, The Electronic Word: Democracy, Technology, and the Arts (Chicago: University of Chicago Press, 1993), 80. ↩
- Ibid., 81 (italics original). ↩
- See Adrienne L. Massanari, ‘Designing for Imaginary Friends: Information Architecture, Personas and the Politics of User-centered Design,’ New Media & Society 12, no. 3 (2010): 401-16, doi:10.1177/1461444809346722. ↩
- This vein of research parallels work on the rhetoric of science, and resonates more broadly with social constructionist approaches to science and technology—for example, see Ronald Kline and Trevor Pinch, ‘Users as Agents of Technological Change: The Social Construction of the Automobile in the Rural United States,’ Technology and Culture 37, no. 4 (1996): 763-95, doi:10.2307/3107097. ↩
- For example, see Barbara Warnick and David S. Heinemen, Rhetoric Online: The Politics of New Media, 2nd ed. (New York: Peter Lang, 2012). ↩
- Lanham, The Electronic Word, 81. ↩
- Ibid. ↩