NICK SEAVER
Tufts University, USA

Abstract

This article examines mouse jigglers—devices that simulate user activity to prevent computers from sleeping—as a generative entry point into the study of media stimulation. Attending to their form and function, the article traces the semiotic and operatory chains within which these devices are inserted. Mouse jigglers intervene in technical systems that aim to render attention measurable and manageable, subverting the sensemaking activities of other actors who depend on conventional links between attention and its signifiers. In these settings, attention is best understood not as an internal human mental state, but as an object distributed across, and constructed by, sociotechnical systems. The use of mouse jigglers is a peculiar consequence of the role of attention in contemporary media environments: it is at once central and underdefined, manifested through proxies that are inevitably vulnerable to manipulation. By interfering in these operations, mouse jigglers reveal the infrastructural logics of legibility at work in contemporary attention tracking software.

Keywords

attention, operational images, optical mice, workplace surveillance

For the official version of record, see here:
Seaver, N. (2026). Keep Computer Awake: Mouse Jigglers and the Technicity of Attention. Media Theory, 10(1), 205–230. https://doi.org/10.70064/mt.v10i1.1373

Keep Computer Awake: Mouse Jigglers and the Technicity of Attention

On YouTube, there exists a strange genre of video: hours of black and white geometrical patterns—grids, diamonds, mazes—which slowly drift across the screen in short, repeating loops (Fig. 1). These videos are high contrast, hypnotic and dazzling, and many of them have hundreds of thousands of views. But on a platform full of psychotechnical content—aids for sleep, study, and other desired mental states—these videos are not made for human eyes. They are a kind of cinematic form designed to be viewed by the sensor on the bottom of an optical computer mouse.

They have titles like “Mouse Jiggler | Mover | Keep Computer Awake” or “MOUSE JIGGLLER! 2:30:00 Hours, move your mouse keep your optical mouse alive while you are away!” The screener of these videos is intended to play them on a smart phone or tablet, placing their mouse on top of the glass. The constantly shifting pattern registers as movement to the mouse, and the motion of the cursor on screen keeps the user’s computer active. Some commenters suggest that their purpose is to prevent a computer from going to sleep during long processes like video rendering; others dubiously claim that the videos can enhance or test mouse “performance.”

But some of the videos and many of the comments point toward a more typical purpose: maintaining an “active” status on workplace communication software like Slack or Microsoft Teams, so that if your boss virtually checks in on you, it looks like you’re paying attention.

If you keep searching for “mouse jiggler” on YouTube, you will find even more ways to feign attention. There are a variety of DIY contraptions designed to stimulate optical mouse sensors. In one, a small, solar-powered “dancing flower” desk toy moves its plastic leaves back and forth; taped to one of these leaves, a piece of paper brushes regularly across the sensor of an upturned mouse. In others, mice are fastened to oscillating fans or set on top of analog wristwatches.

Some videos review dedicated hardware jigglers: these typically consist of a pill-shaped platform with an embedded rotating disk. Their user places a mouse on top and the disk rotates at intervals beneath the sensor, keeping the computer awake indefinitely. Small USB keys present themselves to the computer as mice, producing erratic movement paths when plugged in. Dedicated software programs with stimulant names like “Caffeine” or “Amphetamine” emulate hardware devices, providing just enough input to stop a computer from going to sleep.

Stimulating interfaces

Mouse jigglers tickle computers into wakefulness, simulating attentive users by stimulating the interface. In doing so, they invert a common understanding of the relationship between interfaces and attention. The usual idea is that contemporary interfaces (and the algorithmic apparatuses behind them) stimulate their users into attentiveness, producing addled, hypervigilant attentional subjects. The users of mouse jigglers stimulate computers, rather than the other way around. This article contends that mouse jigglers, despite their rather abject status (an abjection indexed by the term that has become their generic name), offer a unique, and therefore productive, entry point into thinking about how attention is managed and measured today. Attending closely to mouse jigglers’ form and function—as devices meant to intervene in the chains of operations and signification that flow in both directions through interfaces—offers a surprising amount of conceptual leverage.

Bernard Geoghegan (2019: 86) provides an exemplary analysis of the way that interfaces come to shape users’ attention, in his genealogy of “the martial aesthetics of the attention economy.” Contemporary screen interfaces originated in military settings, where their users were expected to remain vigilant through long, monotonous shifts. Design features common to today’s devices—like visual and auditory notifications or real-time representations of the activity of others—find precedents in military screens, whose designers produced a “standard stock of techniques for leveraging users into a state of constant alert” (Geoghegan, 2019: 89). Through the details of their design and the social expectations surrounding their deployment, screen interfaces came to engender the kind of agitated attentional subject we currently associate with excessive computer use. Contemporary viewers of screens inherit this familiar style of attention from an unfamiliar context, molded by an agonistic interface.

In an analysis of social networking sites, Taina Bucher (2012: 4) has argued that we should understand attention not as a narrowly cognitive affair—as a function of individual brains, impinged on from without by technology—but rather “as an emerging property of sociotechnical relations.” This “technicity of attention,” as she terms it, takes seriously critical arguments about the consequences of technologies like social media for our attentional capacities (e.g. Stiegler, 2010; Terranova, 2012), but reframes the problem: these technologies can alter our capacity to attend not because they abuse it from the outside, but because attention itself manifests through interactions “between the human and their technical supports” (Crogan and Kennedy, 2009: 109, cited in Bucher, 2012: 4). Consequently, what attention is and how it works are malleable, susceptible to shifts in the techniques and technologies that mediate between humans and their world.[i]

To build on Bucher, we might claim that attention itself is an interface concept: it describes an inevitable mediation between subjects and their environments. Jonathan Crary (1999: 21) suggests that, in the late-19^th century, attention came to assume “a central and formative role in accounts of how a practical or knowable world of objects comes into being for a perceiver.” This understanding of attention introduced a radical instability that persists today: our experiences of the world are mediated by an arbitrary, subjective, and malleable interface—our attention. Sebastian Watzl (2023), a contemporary philosopher of attention, theorizes it as a “priority structure”: attention organizes mental representations according to a variety of motivational influences (which might include one’s intentions, but could also include other factors, like a history of environmental stimulation that has trained us to attend in certain ways). Attention names the organizing principle of this prioritizing interface between perceiving subjects and the world they act in and reflect upon.

“Attention” is a rather capacious concept, even in technical usage (Taylor, 2023), and ordinary users of the concept do not find it odd or incorrect to speak of brainless entities like robots, surveillance cameras, publics, or corporations paying attention.[ii] One could idiomatically describe workplace surveillance software as “attending” to employees, whether or not an attentive human is watching. Indeed, Watzl argues that his theory of attention is not necessarily constrained to “human-specific representational or neuronal systems,” but might be applied to “diverse cognitive systems” (2023: 5), potentially including non-humans (like birds or AI). Bridging Bucher and Watzl, we might consider the existence of supraindividual, distributed, sociotechnical kinds of attention—structures of selection, exclusion, and prioritization that extend beyond the individual mind and subsume it. Thinking about attention in this way, in some cases, is mostly a shift in emphasis—from saying that screens engender certain styles of attentional subjecthood in users to saying that the complex of screen, user, algorithm, and so on attends in a certain way. But moving from attention-bearing users to attention-producing sociotechnical systems allows us to approach the technical elements of these complex arrangements as components of attention in their own right, not only as outside influences on it.[iii]

In accounts of attention’s technicity like Bucher’s, the technical side of the sociotechnical situation is dominated by powerful institutional actors like algorithms, interfaces, corporations, and so on. As Bucher notes, platforms like Facebook treat their users’ attention as something to be managed and monetized; it is known through various metrics and governed by algorithmic systems that anticipate and contour user behavior. This commonly remarked-upon feature of the online attention economy has spread into the workplace, joining up with a longer history of efforts to manage and surveil workers’ attention (Monahan, 2016; see, e.g., Ajunwa, 2023). Even in accounts insistent that users ultimately retain agency and the ability to resist these systems, the constraining and enabling power of large technical infrastructures nevertheless sets the stage and dominates the interaction.

Mouse jigglers, being technical devices mobilized by users, provide an example of something rarer: technological interventions aimed at manipulating the dominating power, rather than the other way around. We are broadly familiar with top-down efforts to command the attention of computer users, but attention is also worked on from the other end. Worried that their workers are not paying enough attention, bosses use availability status as a proxy for engagement; worried that their bosses are paying too much attention to them, workers rely on elaborate technical contraptions to obscure their coming and going from the computer. Attention as we know it today emerges in this sociotechnical interface—a site of contestation and flux.

To study this situation, then, we should start not at the scale of algorithms and surveillance platforms, but somewhere much smaller: in the space between the mouse and the desk.

Seeing like a mouse

How does an optical mouse work, anyway?

The optical mouse was invented multiple times—twice in December 1980, at Xerox PARC and at MIT (Goy, 1988: 221), and once more at Agilent Technologies, a spin-off of Hewlett-Packard, nearly 20 years later.[i] The general principle across these various devices was the same: the mouse contained an optical sensor, which captured images from beneath it, analyzing changes in those images over time to infer movement.[ii]

The earliest optical mice worked by detecting patterns printed on custom mousepads (the Xerox version came with an actual pad of marked paper, to be consumed over time). In an oral history interview with the Computer History Museum, Gary Gordon (2017), who led the invention of the Agilent optical mouse, described how he had imagined a mouse that might work through optical tracking of the desk surface itself, with no need for an intervening pad. Unfortunately, he lamented, “my desk surface was Formica, and it looked really smooth, and so I was kind of discouraged.” If only, he suggested, “the whole world used wood desks,” with plainly visible patterning, then imaging the desktop might work.

But later on, as Gordon described it, he had an epiphany: “For some reason I scratched the Formica with my finger, and could hear this vibrating sound, and realized that the surface was textured. And then I lit it obliquely with a flashlight, and under a magnifier, you could see all these hills and valleys. It looked like you were in an airplane flying at sunset over hilly terrain.” What had seemed a blank, synthetic surface was revealed by the scratching of a fingernail or a raking light to be an intricately detailed landscape (Fig. 2). This landscape could be imaged, leading to a mouse designed to “navigate upon almost any arbitrary surface” (Gordon et al., 2002).

Gordon often referred to this discovery of the desktop’s hidden topography as the crucial enabler of Agilent’s invention, which was soon licensed to Microsoft, becoming the foundation of most optical mice sold ever since.[i] He was inspired, he said, by his prior experience as an officer on a US Navy aircraft carrier and in military tracking. Around the invention of the Agilent mouse, he was also working on a “flying mouse”—a device that would use the same principles of optical tracking to recognize movement in three dimensions—though it never came to fruition.[ii]

Those principles of optical tracking were derived from earlier work on military aviation—in particular, the study of “optical flow,” initiated by the psychologist J.J. Gibson. In his book The Perception of the Visual World (1950), Gibson theorized about how pilots manage to maintain a sense of stability, despite the sometimes wild motion of both aircrafts and eyeballs, by attending to the texture of the ground. “The flow of the terrain” (Gibson, 1950: 124) over the retina is interpreted as movement—like looking out the window of a car on the highway, which provides evidence of motion despite the lack of other sensory input to corroborate it (Fig. 3). A pilot looking up at a cloudless sky “gets the impression of floating in empty space,” like an optical mouse flipped on its back, “even though he may be objectively moving at 200 miles per hour” (Gibson, 1950: 124).

Texture, in Gibson and Gordon, gives an observer optical traction against a surface, making the detection of movement possible. Gordon’s mouse imaged a roughly two-millimeter–square portion of the desk (or, as one Microsoft press release suggested, a seat cushion or pant leg), capturing about 1000 pixels of visual information every thousandth of a second. Shifting successive images over each other and measuring their correlation would indicate which direction the mouse had likely moved.

Whether we believe Gordon’s account of epiphany, the transformation of the modern desk from a blank space (like the disorienting sky above an airplane) to a textured terrain (like the landscape below it) is a crucial moment in the history of computer interaction. Desks, as Shannon Mattern (2014) has argued, are “infrastructures that we so typically work on without looking at” (emphasis in original); the Agilent mouse inaugurated a strange era of looking at the desk, distributed across hundreds of millions of devices, a few millimeters at a time.[i]

We can think of the mouse jiggler, in this lineage, as a very modest flight simulator. It inverts the roles of mouse and surface, keeping the mouse in one place while shifting the ground beneath it, producing evidence of motion that the mouse duly transmits to a computer that doesn’t know any better than to believe its “eye.”[ii]

Operational images

The images captured by the optical mouse sensor are, in Harun Farocki’s (2004: 17) terms, “operational”: they “do not represent an object, but rather are part of an operation.” The optical data gathered by the sensor are not intended for human audiences, not displayed anywhere, and, in fact, do not even escape from below the hand, being processed on dedicated hardware within the mouse itself. As Jussi Parikka elaborates in his book on operational images (2023: 31), operationality implies “the primacy of action and function” over the production of images “to be seen and interpreted for meanings.” We might say that operational images are essentially technical, playing a role in what anthropologists of technology would call the “operatory chain” (Leroi-Gourhan, 1993; Coupaye, 2022; see Parikka, 2023: 35–36) —the series of transformative actions that constitutes any technical activity.[i]

Farocki originated the idea of the operational image in his artistic work on aerial drone photography, used by militaries to identify missile targets. These photographs are taken less to be looked at than to be used, as inputs to military decision-making. As Farocki notes, they are typically discarded once their function has been served. In an earlier essay on the “instrumental image,” which anticipated the operational, Allan Sekula (1975) analyzed another set of aerial reconnaissance images, orchestrated by the photographer Edward Steichen for military use during WWI but curated into an artistic collection afterwards. These images play on the troubled boundary between the “functional” and the “aesthetic.” What makes them “instrumental” in their original context, Sekula contends, is a rationalized interpretive practice that is relentlessly focused on the indexicality of signs within the image: “Is that a machine gun, or a stump?” The military viewer, unlike later artistic audiences, was constrained to a narrowly analytic, indexical interpretation—a kind of visual attention that could be imagined to simply see what was there.

We might say, then, that what makes an image “operational” (or instrumental) is a socially organized expectation of controlled attention. The military viewer is not supposed to think about what the images remind him of, their formal aesthetic qualities, or anything like that. Instead, he is meant to narrow his focus, to observe in a technical manner, trying to establish the concrete indexical referents of the images. The interpretive surfeit we might associate with aesthetic viewership is curtailed. There is, as we can see here, a connection between indexicality and operationality—a narrowing of the world of interpretive possibility.

More recent work on operational images has turned to situations in which human viewers are absent or displaced to the margins. Machine vision systems (Paglen, 2014) process images “invisually” (MacKenzie and Munster, 2019), and generative AI systems produce pictures without the outward reference to represented objects or perceiving subjects typically associated with vision. Here, we start to see how the category of the operational image pushes against the limits of what we think of as “images.” And, in the literature on operational images, this is associated with a kind of function-first technicity.

The optical mouse, flying across the desktop, fits comfortably within this lineage of aerial photography, though its sensor is so minute we might hesitate to call it a “camera.” While I have focused on the functional nature of operationality here, scholars like Parikka have argued that “operations” should be understood more broadly. Operationality is not just a matter of images being used within operatory chains, but also of their role within the history of operations research as a field, related to “questions of logistics in massive technical systems that work through the ability to abstract and optimize” (Parikka, 2023: 31). That is, it is not enough to be technical; when we encounter operational images we should be looking for their role within larger systems of mediation and control.

From the individual user manipulating their mouse, we can zoom out to see how mouse movement is abstracted and analyzed en masse, in the workplace scenarios that occasion the use of mouse jigglers in the first place. Some workers report informal surveillance, in which managers simply note whether their employees are “active” on workplace communication software like Microsoft Teams, looking for “the little green dot” that indicates availability. In other cases, this surveillance is systematized, in software packages (sometimes called “bossware”), that collect data about employee computer activity, including websites visited, applications used, and active time. Bossware platforms like ActivTrak or Teramind aggregate and analyze this data, presenting managers with summaries that quantify their employees’ “focused time,” “productivity,” and so on, in a variety of customizable reports.[ii]

In its conventional use, the optical mouse is part of an operatory chain that links the hand of the user to the movement of the cursor on screen. Bossware surveillance systems are appended to the end of this chain, extending it into new functions. Now, the movement of the mouse triggers a further sequence of technical operations, which stretch out into surveillant apparatuses that take the movement of the mouse across the desk as an index of employee attentiveness.

Apotropaic functions

At the other end of the chain, we find the worker, who may be interested in blocking such managerial attention. Such a worker may find the geometrical videos described above useful for appending to their own end of the operatory chain. Like the operational images generated within the mouse, these videos can be understood as functional more than aesthetic. But unlike sensor data, these videos are not images of anything—they are not indexes of an outside world. Rather, they are synthetic productions, made on computers for a computational audience. They are designed around a vernacular theory of human and machine perception, anticipating both the optical mouse’s sensor and the human eye that will eventually look for an “active” status on Teams or in the logs of Teramind. They anticipate the controlled, indexical interpretation of operationality and are not meant to be interpreted for wider resonances.

But we can interpret them anyway, and when we do, we might spot a similarity between their dazzling patterns and a variety of repeating geometrical motifs that recur across the ethnological record, such as Greek and Roman labyrinths, Celtic knotwork, South Asian kolams, and the intricate designs of Trobriand canoe prows (Fig. 4).[i] In his book Art and Agency (1998), Alfred Gell offers a catalog and analysis of such patterns, providing an account of their function. European modernist thinkers like Adolf Loos (2019) disdained patterns like these as nonfunctional ornament, to be discarded by “mature” civilizations in favor of designs that hewed closer to function. Loos used the intricate tattooing found in Papua New Guinea as an example of the tendency of “less evolved” races toward nonfunctional ornament.[ii] Against such explanations, Gell argues that intricate ornamentation is in fact functional.

Ornamentation is a psychological technology, he suggests, as is art more generally, designed to produce particular mental states in observers (Gell, 1992). Critics of ornament are inadvertently correct to call it “tacky,” he writes, because the experience of actually engaging with such patterns is adhesive: “once one submits to the allure of the pattern, one is liable to become hooked, or stuck, in it” (Gell, 1998: 82). Gell elaborates a theory by which human viewers find such intricate patterns enchanting, imagining the agency that can command the creation of a design that is so complex that the means of its construction cannot be easily adduced. But frequently, the intended audiences for such patterns are not human.

Much geometric art, Gell notes, is apotropaic: it is designed to protect its bearer or maker against its audience or recipient, and that audience is frequently demonic. Local accounts of Tamil kolams or Celtic knotwork suggest that their complexity is irresistibly tantalizing to evil spirits, drawing them away from the people they might otherwise bedevil with their attention. “Apotropaic patterns are demon-traps,” Gell writes (1998: 84), “demonic fly-paper, in which demons become hopelessly stuck, and are thus rendered helpless.” This mechanism, it is worth emphasizing, is thoroughly attentional: the goal is to redirect the attention of an enemy, to funnel it toward a pattern that is essentially an attention-sink, a pattern so complex and repetitive that it is challenging to turn away. Such traps produce stasis from intricate local motion, capturing their prey in a stimulating cognitive micro-world that, from another frame of reference, looks like nothing is happening.[iii] Gell rejects the distinction between aesthetic and functional images, arguing that even the most ornamental of patterns do, in fact, serve a function.

At first, this may seem to pose a challenge to the theory of operational images, which, as I’ve described, hinges on the identification of a category of distinctively functional, non-aesthetic image making and interpretive practice. But Gell is not concerned with such a distinction (for better or worse), and he is happy to subsume “aesthetic” interactions under a broader umbrella that also contains functional ones. Thus, in one memorable example from Art and Agency (1998: 66), Gell suggests that his theory of artworks’ function might encompass both someone looking at a holy icon while appreciating its symbolism and someone kissing a holy icon and being cured of their rheumatism.

Thinking about video jigglers as a genre of apotropaic imagery also provides an entry point into understanding the aesthetics of hardware mouse jigglers, with their textured rotating disks. Rotation is a popular cultural technique for producing repeating patterns in decorative art (Gell, 1998: 170; Eggelhöfer, 2018), which renders simple motifs into designs of entrancing complexity. The hardware mouse jiggler takes surface texture—that essential element of optical flow, which grounds perceiving agents in their environment according to Gibson—and manipulates it into controlled local motion. Anticipating the mouse’s sensor, the hardware jiggler is designed to hold the mouse in place (often quite literally, with a raised rim around the edge of the device) and to occupy its modest sensorium with irresistibly patterned inputs. Though we might be inclined to distinguish between videos and hardware, both of these types of mouse jiggler are technologies of visual display.

So, the felicitous visual similarity between mouse jiggler videos and other forms of demonic flypaper from around the world is not merely superficial. Like the kolam or the labyrinth, these patterns are apotropaic and operational, mobilized to defend their bearers against the attention of hostile agents. And like other apotropaic practices, they evince a model of their audience’s perceptual and attentive tendencies, anticipating particular kinds of viewer. Two forms of attention collide in the mouse jiggler, which aims to interrupt the attention of surveillant bosses and to simulate the attention of active workers.

---

[i] To this ethnological set we might add the striped “dazzle camouflage” applied to ships in WWI. See Jessa Lingel’s “Dazzle Camouflage as Queer Counter Conduct” (2021) for a discussion of other, similar forms of resistance to surveillance in relation to queer theory. Lingel draws helpfully on M. Murphy’s (2012: 183) theorizing of “counter-conduct” as an alternative to the often romanticized status of “resistance”: where resistance may be valorized and imagined as an alternative to hegemony, counter-conduct is “immanent with and animated by hegemonic formations”—as when an employee uses a mouse jiggler not to contest logics of workplace surveillance, but to find a way to work within it.

The semiotics of deception

The links in the operatory chain we have been tracing are not only material, but semiotic, articulated by interpretive operations that posit relationships between signs and objects. To think about how these interpretive connections are made, I will briefly turn to a Peircean framework, as elaborated by the linguistic anthropologist Paul Kockelman (2005; 2024). In that framework (represented in Fig. 5), agents take signs to stand for objects, instigating interpretants in response to them. So, for example, you (the agent) might hear the sound of a smoke detector going off in the kitchen (the sign), understand that your dinner has caught fire in the oven (the object), and go to investigate (the interpretant).[i] Or in the case of the optical mouse: the mouse (agent) detects a shift in the texture beneath it (sign), which it interprets as movement (object), sending a signal to the computer to move the cursor on the screen (interpretant). In this framework, agents need not be human and interpretants need not be intentional.

It is possible, you may note, to zoom in on various parts of this process and find more signifying nested at smaller scales. Where I say that “the mouse” senses a sign, we could instead describe that action in terms of semiotic relations among the mouse’s components. The optical sensor “interprets” photons into arrays of bits, which it passes on to a chip that interprets the relationship between successive arrays as a certain amount of motion, which it passes on to the computer as an input command, and so on. These signifying situations are not only nested but enchained, with the interpretants of one process serving as signs in the next. Following the chain from the mouse: the movement of the cursor on screen (formerly interpretant, now sign) may be taken as an index of user activity (object) by Microsoft Teams (agent), which will set the user’s status to “active” (interpretant). That user’s manager, looking through their list of employees in Teams, may see the green dot next to the employee’s name, understand it to indicate that they’re working, and develop a sense that they are diligent.

Conventional operatory chains take technical artifacts (like stone hand-axes discovered in archaeological sites) and analytically locate them within longer processes (like identifying candidate stones, chipping them against harder materials, and then testing their sharpness against softer ones). But, as Ludovic Coupaye (2022: 46, emphasis in original) suggests, “not all technical activities have a visible or empirical material result.” While our prototypical technologies may be essentially material, technical activity can be defined (following Mauss, 1935) as “effective” and “traditional”—that is, action oriented toward some goal (whether it succeeds, is well-informed, or works efficiently), which is emplaced within a broader context where techniques are shared and learned (not necessarily “traditional” in the sense of claiming distant origins). So, technical activity includes material action like making a hand axe, but also sending a smoke signal, targeting a missile, taking a photograph, surveilling an employee, or using a mouse jiggler.[i]

Signs do not signify in isolation. To link signs with objects—to understand that smoke means fire, optical flow means motion, or “active” means attentive—requires interpretive grounds (which Kockelman [2024: 8] also calls “values”). This grounding can take many forms: the optical mouse is grounded in the principles of optical flow encoded into its design; Microsoft Teams interprets mouse input as evidence of user activity because it has been programmed to do so; managers interpret “active” statuses as evidence that their employees are working because they have been encouraged to by the design of the interface and, perhaps, by their own bosses. Looking back to Sekula’s account of instrumental images, the disciplined indexicality of military observers is also semiotically grounding in this sense, determining how relations between signs and objects should be assembled.

This grounding seems, especially in the use of images deemed “operational,” rather technical in nature—that is, it is not only an interpretive convention that causes the mouse to treat optical flow as evidence of motion, but the physical configuration of microchips and sensor arrays. In actor-network theory terms, we might say that this interpretive convention has been “made durable” by rendering it in hardware (Latour, 1990). Analyzing semiotic and operatory chains together, we can borrow bits of one theory to inform the other: as the links in the semiotic chain are grounded in broader contexts that establish plausible connections between signs and objects, so too are links in the operatory chain informed by shared ideas about what is possible, reasonable, and desirable.

Mouse jigglers undermine these conventions, so that computers perceive activity and managers perceive attention where they are in fact absent. As Kockelman (2024: 10) notes, “a key feature of many semiotic processes is the fact that the agent only learns about the object through the sign.” There is, in many signifying processes, a layer of obscurity between signs and their objects. Managers cannot access their employees’ attention directly, so must rely on intermediary signs, and the same is true of all efforts to identify attention, which is ultimately an imputed internal state. Teachers, social media marketers, or newspaper executives rely on signs taken to indicate the presence of attention, like posture, clicks on online advertisements, or time spent on particular web pages. It is simply impossible to engage attention “directly,” in an unmediated way.[ii]

In these cases, the signs that indicate attention are taken to be indexical—that is, they are understood to stand for their object not in some arbitrary (i.e., symbolic) or mimetic (i.e., iconic) way, but rather by virtue of being directly connected to them. User input activity is presumed to be a natural component of attentive computer use.[iii] Deceptive interventions like the mouse jiggler undermine that presumed indexicality, effectively short-circuiting the semiotic processes of other agents. The obscurity of the object, which is an obstacle for would-be attention detectors, is an opportunity for attention simulators: a model of other agents’ semiotic grounds makes it possible to intervene in their sensemaking practices.[iv] Workers use mouse jigglers to trick their bosses, drawing on an understanding they have about what their bosses care about and how they know about it.

Attention has been replaced with something roughly attention-shaped, and the system keeps on functioning despite the absent object.[v] As a result, we can see that, despite the central role played by the projected object, it is not really the object that causes the interpretant, but the sign (Fig. 6). Attention can be removed and the system will proceed as a simulacrum of itself. That is, until one’s manager begins to suspect, through other signifying chains—like the failure to respond to messages, the missing of work meetings, or the absence of work outputs—that an employee is not, in fact, working attentively at their desk.

The ends of attention

The possibility of being fooled is not a surprise to the designers of bossware platforms, and companies like Teramind have upgraded their software in attempts to address it. On the company blog, Teramind shares posts with titles like “What is Activity Falsification, and Why Should You Care?” and “How to Easily Detect Employees Using Mouse Jigglers” (Joyce, 2024; Parsons, 2024). These content marketing posts (designed to attract traffic to the company website when people search for phrases like “are mouse jigglers detectable”) advertise Teramind’s use of machine learning to model employees’ typical input behaviors. The company claims that “roughly 7% of employees actively ‘fake’ digital work activity” (Parsons, 2024), which it discovered while developing a tool for validating worker identity—modeling how employees behave so that dramatic changes, which might indicate that a different person is using the computer or that automated tools are in use, can be detected.

These are marketing claims, so we should be skeptical about taking them at face value as evidence of managerial capacities, but they are at least indicative of managerial desires and form part of the background against which digitally savvy workers think about how mouse jigglers might be detected. The posts are vague about how Teramind’s system works, only indicating that their “proprietary models […] can detect patterns that deviate from the user’s baseline behaviors for input speed, consistency, and variation” as well as “inorganic movement” (Parsons, 2024). This vagueness is typical of the prosaic, marketing-driven applications of machine learning that Jonathan Sterne (2025) has called “mid AI,” but it may also be related to a desire to keep particular criteria secret from those who might use them to build more effective attention simulators.

In online discussions of detectability (which are very prevalent in fora where mouse jigglers are discussed, like subreddits, Amazon reviews, and YouTube comments), posters speculate on what might be necessary to head off these automatic schemes. Most hardware jigglers have randomization controls built in, which change the direction of rotation at irregular intervals. More elaborate software systems, which claim to circumvent employers’ ability to detect and block the use of unauthorized programs, now incorporate a variety of input simulation techniques, switching among applications, scrolling through documents, and emulating keypresses. LazyWork, a subscription-based activity simulation program, offers an “Elite” tier that costs $49.99 USD per month and also claims to solve “attention popups”—explicit tests of user attentiveness used by some employers—in addition to simulating cursor movement, clicks, and so on.

There is, in principle, no end to this arms race between attention detectors and attention simulators, and some online commenters fantasize about a future in which the automated attention measurement of platforms like Teramind is forever met by the automated attention simulation of mouse jigglers, leaving the humans on either side of the interface free to get on with their lives.[i] In this fantasy, attention has been abstracted entirely away from human subjects, transformed into a semiotic relation between machines. Here, attention’s technicity is taken to an extreme, with two mutually attentive machines locked together indefinitely. To return to where we began, we might consider these bound machines as technical counterparts to the addled, hypervigilant computer user, stimulated into pathological attentiveness, caught by the designs of media interfaces meant to transfix them. Here, it is the computers, stimulated into wakefulness and set into antagonistic relation with other machines, that have been caught, drawn in by the irresistible designs of operatory and semiotic chains along which human and non-human attention travels, and through which they are formed. Once the machines have distracted each other, perhaps we might attend to something else.

References

Ajunwa, I. (2023) The Quantified Worker: Law and Technology in the Modern Workplace. Cambridge: Cambridge University Press. Available at: https://doi.org/10.1017/9781316888681.

Bucher, T. (2012) ‘A technicity of attention: How software “makes sense”’, Culture Machine, 13, pp. 1–13.

Chun, W.H.K. (2011) Programmed Visions: Software and Memory. Cambridge, MA: The MIT Press.

Coupaye, L. (2022) ‘Making “Technology” Visible: Technical Activities and the Chaîne Opératoire’, in Bruun, M.H. et al. (eds.) The Palgrave Handbook of the Anthropology of Technology. Singapore: Springer Nature Singapore, pp. 37–60. Available at: https://doi.org/10.1007/978-981-16-7084-8_2.

Crary, J. (1999) Suspensions of Perception: Attention, Spectacle, and Modern Culture. Cambridge, Mass: MIT Press.

Crogan, P. and Kennedy, H. (2009) ‘Technologies Between Games and Culture’, Games and Culture, 4(2), pp. 107–114. Available at: https://doi.org/10.1177/1555412008325482.

Eggelhöfer, F. (2018) ‘Shifting, Rotating, Mirroring — Lena Bergner’s Minutes of Paul Klee’s Classes’, Bauhaus Imaginista. Translated by K. Hoffmann, 3 August. Available at: http://www.bauhaus-imaginista.org/articles/1489/shifting-rotating-mirroring (Accessed: 17 May 2025).

Farocki, H. (2004) ‘Phantom Images’, Public [Preprint]. Available at: https://public.journals.yorku.ca/index.php/public/article/view/30354 (Accessed: 4 May 2025).

Forsythe, S. (2023) ‘Adversarial Abduction: The Logic of Detection and Deception’, in Magnani, L. (ed.) Handbook of Abductive Cognition. Cham: Springer International Publishing, pp. 1465–1487. Available at: https://doi.org/10.1007/978-3-031-10135-9_32.

Gell, A. (1992) ‘The technology of enchantment and the enchantment of technology’, in Coote, J. and Shelton, A. (eds.) Anthropology, Art and Aesthetics. Oxford: Oxford University Press, pp. 40–63.

Gell, A. (1998) Art and Agency: An Anthropological Theory. Oxford: Clarendon Press.

Geoghegan, B.D. (2019) ‘An Ecology of Operations: Vigilance, Radar, and the Birth of the Computer Screen’, Representations, 147(1), pp. 59–95. Available at: https://doi.org/10.1525/rep.2019.147.1.59.

Gibson, J.J. (1950) The Perception of the Visual World. Boston, MA: Houghton Mifflin. Available at: http://archive.org/details/perceptionofvisu00jame (Accessed: 25 September 2024).

Giedion, S. (1948) Mechanization Takes Command. New York: Oxford University Press.

Gordon, G. (2017) ‘Oral history of Gary Gordon’. Interviewed by Gunter Steinbach, 1 March. Mountain View, CA: Computer History Museum. Available at: https://www.computerhistory.org/collections/catalog/102706894/ (Accessed: 11 June 2026).

Gordon, G.B. et al. (2002) ‘Seeing eye mouse for a computer system’. Available at: https://patents.google.com/patent/US6433780B1/en (Accessed: 13 May 2025).

Goy, C. (1988) ‘Mice’, in Sherr, S. (ed.) Input Devices. San Diego, CA: Academic Press, pp. 219–250.

Joyce, A. (2024) ‘How to Easily Detect Employees Using Mouse Jigglers’, Teramind, 25 July. Available at: https://www.teramind.co/blog/how-to-detect-a-mouse-jiggler/ (Accessed: 27 May 2025).

Kockelman, P. (2005) ‘The semiotic stance’, Semiotica, 2005(157), pp. 233–304. Available at: https://doi.org/10.1515/semi.2005.2005.157.1-4.233.

Kockelman, P. (2024) Last Words: Large Language Models and the AI Apocalypse. Chicago, IL: Prickly Paradigm Press.

Latour, B. (1990) ‘Technology is Society Made Durable’, The Sociological Review, 38(1_suppl), pp. 103–131. Available at: https://doi.org/10.1111/j.1467-954X.1990.tb03350.x.

Leroi-Gourhan, A. (1993) Gesture and Speech. Translated by A. Bostock Berger. Cambridge, Mass: MIT Press. Available at: http://archive.org/details/leroi-gourhan-andre-gesture-and-speech (Accessed: 15 May 2025).

Lingel, J. (2021) ‘Dazzle Camouflage as Queer Counter Conduct’, European Journal of Cultural Studies, 24(5), pp. 1107–1124. Available at: https://doi.org/10.1177/1367549420902805.

Loos, A. (2019) Ornament and Crime. Translated by S. Whiteside. London: Penguin Classics.

MacKenzie, A. and Munster, A. (2019) ‘Platform Seeing: Image Ensembles and Their Invisualities’, Theory, Culture & Society, 36(5), pp. 3–22. Available at: https://doi.org/10.1177/0263276419847508.

Martin, E. (2021) Experiments of the Mind: From the Cognitive Psychology Lab to the World of Facebook and Twitter. Princeton, NJ: Princeton University Press.

Mattern, S. (2014) ‘Intellectual Furnishings’. Working paper presented at the Graduate Institute of Design, Ethnography, and Social Thought Bi-Weekly Seminar, 7 November. Available at: https://medium.com/@shannonmattern/intellectual-furnishings-e2076cf5f2de (Accessed: 11 June 2026).

Mauss, M. (1935) ‘Techniques of the body’, Economy and Society, 2(1), pp. 70–88. Available at: https://doi.org/10.1080/03085147300000003.

Monahan, T. (2015) ‘The Right to Hide? Anti-Surveillance Camouflage and the Aestheticization of Resistance’, Communication and Critical/Cultural Studies, 12(2), pp. 159–178. Available at: https://doi.org/10.1080/14791420.2015.1006646.

Monahan, T. (2016) ‘Built to lie: Investigating technologies of deception, surveillance, and control’, The Information Society, 32(4), pp. 229–240. Available at: https://doi.org/10.1080/01972243.2016.1177765.

Murphy, M. (2012) Seizing the Means of Reproduction: Entanglements of Feminism, Health, and Technoscience. Durham, NC: Duke University Press.

Ogden, M. (2023) ‘Package Wars and Mouse Movers: On the Media Escalation of Remote-Work during the COVID-19 Global Pandemic’, Culture, Theory and Critique, 64(3), pp. 356–366. Available at: https://doi.org/10.1080/14735784.2023.2299468.

Paglen, T. (2014) ‘Operational Images’, e-flux [Preprint], (59). Available at: https://www.e-flux.com/journal/59/61130/operational-images/ (Accessed: 17 June 2021).

Parikka, J. (2023) Operational Images: From the Visual to the Invisual. Minneapolis [Minnesota]: University of Minnesota Press.

Parsons, L. (2024) ‘What is Activity Falsification, and Why Should You Care?’, Teramind, 15 February. Available at: https://www.teramind.co/blog/what-is-activity-falsification-and-why-should-you-care/ (Accessed: 27 May 2025).

Pedersen, M.A., Albris, K. and Seaver, N. (2021) ‘The Political Economy of Attention’, Annual Review of Anthropology, 50(1), pp. 309–325. Available at: https://doi.org/10.1146/annurev-anthro-101819-110356.

Schor, N. (1987) ‘Introduction’, in Reading in Detail: Aesthetics and the Feminine. New York: Methuen.

Seaver, N. (2019) ‘Captivating Algorithms: Recommender Systems as Traps’, Journal of Material Culture, 24(4), pp. 421–436. Available at: https://doi.org/10.1177/1359183518820366.

Seaver, N. (2023) ‘Attention Is All You Need: Humans and Computers in the Time of Neural Networks’, in Burnett, D.G. and Smith, J.E.H. (eds.) Scenes of Attention: Essays on Mind, Time, and the Senses. New York, NY: Columbia University Press, pp. 230–248. Available at: https://www.degruyterbrill.com/document/doi/10.7312/burn21118-012/html (Accessed: 14 May 2025).

Sekula, A. (1975) ‘The Instrumental Image: Steichen at War’, Artforum, 1 December. Available at: https://www.artforum.com/features/the-instrumental-image-steichen-at-war-209590/ (Accessed: 9 May 2025).

Sterne, J. (2025) The Sound of AI [Lecture]. 19 February, Radcliffe Institute for Advanced Study, Cambridge, MA.

Stiegler, B. (2010) Taking Care of Youth and the Generations. Translated by S. Barker. Redwood City, CA: Stanford University Press.

Taylor, H. (2023) ‘Attention as a Patchwork Concept’, European Journal for Philosophy of Science, 13(3), article 36. Available at: https://doi.org/10.1007/s13194-023-00538-5.

Terranova, T. (2012) ‘Attention, economy and the brain’, Culture Machine, 13(1), pp. 1–19.

Watzl, S. (2023) ‘What attention is. The priority structure account’, WIREs Cognitive Science, 14(1): e1632. Available at: https://doi.org/10.1002/wcs.1632.

Notes

[1] On the techniques and technologies of attention, see Pedersen, Albris and Seaver, 2021.

[2] See also, for instance, the “attention mechanisms” underlying the current generation of generative AI systems (Seaver, 2023).

[3] In Watzl’s forthcoming work, he describes such a setting as a “landscape” in which attention acts, rather than as a component of attentional selection itself.

[4] MIT: US patents 4,364,035 and 4,390,873; Xerox: US patents 4,521,772 and 4,521,773.

[5] For a longer critical history of the mouse, see Rae Bruml Norton’s ongoing project, Complication of the Computer Mouse. Available at: https://itp.nyu.edu/shows/spring2019/category/student/emma-norton/ (Accessed: 20 June 2026).

[6] See, for instance, his 2003 interview with the magazine Fast Company. Available at: https://www.fastcompany.com/1538786/gary-gordon-fast-50-2003 (Accessed: 20 June 2026).

[7] However, Gordon did file a related patent (US6489945B1) for a “method and system for tracking attitude” of such a hand-held pointing device.

[8] Mattern’s essay on intellectual furnishings engages extensively with Siegfried Giedion’s history of the desk and other furniture in Mechanization Takes Command (1948).

[9] The Agilent patent referred to its invention as a “seeing-eye mouse.”

[10] This concept, originated by Leroi-Gourhan in French as the “chaîne opératoire,” goes by a variety of names in English, including operatory chain, operational sequence, operative chain, and their permutations.

[11] Elsewhere, we are undertaking a study of online worker discourse about the use of mouse jigglers under these conditions; see Nick Seaver, Wanjie Li, and Asli Kocak, “Jiggling Against the Machine: Moral Orders of Attention Simulation,” manuscript in progress.

[12] To this ethnological set we might add the striped “dazzle camouflage” applied to ships in WWI. See Jessa Lingel’s “Dazzle Camouflage as Queer Counter Conduct” (2021) for a discussion of other, similar forms of resistance to surveillance in relation to queer theory. Lingel draws helpfully on M. Murphy’s (2012: 183) theorizing of “counter-conduct” as an alternative to the often romanticized status of “resistance”: where resistance may be valorized and imagined as an alternative to hegemony, counter-conduct is “immanent with and animated by hegemonic formations”—as when an employee uses a mouse jiggler not to contest logics of workplace surveillance, but to find a way to work within it.

[13] For more on the erotics and economics of the ornament in Loos, see Naomi Schor’s feminist account of the detail in Reading in Detail (1987: esp. 50–58).

[14] I leave the comparison with a person enchanted by the behavioral design of their smartphone as an exercise for the reader (Seaver, 2019).

[15] It’s also worth noting that the interpretant is not the same as the interpreter—it is the outcome of the sign’s interpretation by the agent.

[16] If one is concerned that such a capacious definition renders most or all human activity potentially “technical” (in a derogatory sense), it is worth bearing in mind Coupaye’s (2022) reminder that the operatory chain is essentially a methodological, analytic, representational tool—a way of tracing and showing “transects” through a heterogeneous, reticulated world. The point is not to set the technical apart from other domains and claim territory for it, but to examine the technical aspect of a wide variety of human behavior.

[17] This is plainly true in mediated settings like remote work, but it is also the case in face-to-face interaction, where a variety of manipulable signs like eye contact, posture, affirmative “mmhmms,” and the like are used to infer and signal an internal mental state. It is also the case in the hyper-controlled technosemiotic settings of psychological experimentation, where much effort is put into establishing conditions that allow “attention” to become evident (see, e.g., Martin, 2021).

[18] For more on the “directness” of computer interface actions, see Wendy Chun’s Programmed Visions (2011).

[19] This is a reversal of Monahan’s (2016) account of workplace surveillance software, which figures the platform itself as deceptive (in that it may be implemented without workers’ knowledge), or at least ethically dubious “spyware.” In another paper on aesthetically striking anti-surveillance camouflage, Monahan (2015: 171) suggests that such techniques (like makeup that confuses facial recognition algorithms) are “hyper-individualized and consumer-oriented” responses to systemic problems; the market for mouse jigglers, which emerged mostly since the publication of his paper, seems to corroborate his analysis.

[20] Samuel Forsythe (2023; and pers. comm.) suggests that deception more generally can be understood as the iconic replacement of indices in situations where mediation prevents other modes of access to the object: swapping out signs taken to directly signify the presence of otherwise undetectable objects with signs that resemble them but lack the direct connection of the index. Examples might include altering photographs, filling a box of “gold” with lead shot, or placing an optical mouse on a rotating disk.

[21] In a recent article, Malcolm Ogden (2023: 365) considers mouse jigglers to be part of a more widespread “media escalation”: “in relatively affluent post-industrial settings, a mundane and media-driven defining and detecting of ‘enemies’ constantly takes place, at work and at home.”

Nick Seaver is an associate professor of Anthropology at Tufts University, where he also directs the program in Science, Technology & Society. He is the author of Computing Taste: Algorithms and the Makers of Music Recommendation (2022), an ethnographic study of music recommender system developers, and co-editor of Towards an Anthropology of Data (2021).

Email: nick.seaver@tufts.edu

Conflicts of interest
None declared

Funding
This research was funded by the Canadian Institute for Advanced Research’s Future Flourishing Program.

Article history
Article submitted: 28/05/25
Date of original decision: 01/09/25
Revised article submitted: 25/09/25
Article accepted: 25/09/25