Is a boundless photographic image still a photograph? Spatially, boundaries define a photo, imposing limits that demarcate its content or subject as this and not that. The edge of a camera’s recording surface, be it CCD or CMOS, negative or positive, wet or dry, circumscribes a limit beyond which no light values are recorded. The frame is thus essential to the medium, defining it as much by what it excludes as what it includes. A photo of a cat on a lap is different from one that includes the person’s face, and the boundary or frame provides the critical tool of composition that makes this determination. Only this much is visible, and no more. So what is an image with no boundary at all?Attempts to move beyond the frame toward boundlessness typically rely on combining multiple images, a process now commonly referred to as “photo stitching.” This allows the final image to exceed the limits of capture by forging the one from the many. As the artist David Hockney once claimed, “if you make a collage using plenty of negatives, you’re actually making a picture with quite a big negative.”1 In the analog era, this might involve cropping or overlapping the images to align their contents, working to conceal or reveal the seams as the photographer saw fit. In the digital era, this is all handled by software. Extending the edge via stitching involves close attention to the contents of each image on either side of the seam, using a procedural algorithm to determine how the images are conjoined. This is as true for analog images as it is for digital ones. Choices must be made about which parts of the constituent photos to save or sacrifice in service of the larger image. Photo-stitching algorithms dictate the rules governing the relation of the parts to the whole. They are, in other words, a kind of politics of the image.Invoking politics here allows us to consider the relationship between the assembly of images in photo stitching and the common register of assembly at work in democratic politics, those collectivities that gather physically to levy a political demand upon the state. In what follows, I argue that these two forms of political assembly—the assembly of images and the assembly of people—bear more than a passing resemblance. They form a sort of recursive loop with one another, where increasingly boundless images reflect our increasingly fluid yet rigidly circumscribed geopolitical moment. To explore this, I will consider the boundless panoramas that define the massively mundane project known as Google Street View and the extent to which the implicit politics at work on the platform reflect or resist the human assemblies that travel within this same space.Rights of assembly, the freedom of citizens to gather publicly for political and other purposes, are long enshrined in Anglo-American case law, appearing first in the Magna Carta in 1215 and then reasserted in the first amendment of the Bill of Rights.2 This mode of political expression has taken various forms in the history of liberal governance, playing an instrumental role at key turning points in American history, including antebellum abolitionism, women’s suffrage, the labor movement, and the civil rights movement of the 1960s. As this lineage demonstrates, the right of assembly might be enshrined within the law, but it is explicitly a tool to challenge and change it, one often wielded by those excluded from its protections.Toward the end of the last century, assembly as a means of organizing and expressing political opposition seemed to wane, to the extent that legal scholar John Inazu described it in 2009 as a “forgotten freedom” that nonetheless sits “at the core of democracy.”3 In the wake of both the Occupy movement and the various social movements referred to (somewhat problematically) as the Arab Spring, assembly reemerged in movements around the world in a newly energized and evolving form.4 Assemblies have appeared in a variety of places and political contexts, from the United States and Europe to South America, the Middle East and Asia, and in established democracies, military dictatorships, and frameworks somewhere in between. Under a range of names that run from descriptive to compelling to playful—Black Lives Matter, the Yellow Vest movement, the Umbrella Movement, the Milk Tea Alliance—gatherings of bodies in the street have once again become a prominent and powerful tool for political engagement.In spite of their heterogeneity, however, these assemblies share several essential characteristics. As is frequently noted in press coverage and academic analysis, these movements maintain a complicated and in some ways antagonistic relationship with mobile technology and social media (at once tools for connective mobilization and for state surveillance).5 While these tools are often credited with enabling the rise of “networked” assemblies, what is perhaps more remarkable is the persistence of embodied participation even as so much political discourse has moved online. Well into the era of platform politics, the collective occupation of physical space remains an essential element for exercising collective power.Further, following Judith Butler, physical presence enables these assemblies to constitute a type of embodied, performative politics.6 In this framework, embodied assembly performs a demand for recognition and a claim to some form of political legitimacy. For Butler, physical co-presence foregrounds the shared precarity that unifies and mobilizes the assembly while grounding their claims to political redress. Appearing in public space entails a degree of vulnerability, a form of physical precarity that mirrors the more amorphous forms of precarity faced by those assembled (economic exploitation, institutional discrimination, structural racism, etc.). Quoting Hannah Arendt, she argues that the appearance in the street of those denied the conditions to live is an embodied assertion of “the right to have rights.”7 In this formulation, the recourse to things like livestreaming and social media during confrontations with police are in fact a product of the shared, physical vulnerability of bodies in the street rather than an instigating force behind their constitution, as some would have it.8 As a corollary, these assemblies also frequently entail the prolonged occupation of space, either tactically or simply out of sheer necessity.Finally, rather than coalescing around institutional hierarchies and centralized leaders, these assemblies evince a strong affinity for decentralized, horizontal procedures of decision-making and power sharing. This tendency manifests itself in what Michael Hardt and Antonio Negri describe as a deep suspicion of traditional leadership, and a latent if undirected need and desire for more radically democratic modes of governance that would obviate the role of traditional sovereignty.9 For Hardt and Negri, this horizontality is both the source of their collective power to destabilize the existing state, but also the Achilles heel of social movements represented by assembly. Hence the need for what they repeatedly refer to as “non-sovereign institutions.” This horizontal decentering of leadership further indicates a centripetal shift from the importance of the individual to the preeminence of the group. As a Thai protester, responding to the arrests of several high-profile figures asserted, “They think arresting the leaders will stop us…It’s no use. We’re all leaders today.”10 This assertion reveals both the individual vulnerability and collective power that individuals face vis-à-vis police and the state, thereby connecting horizontal decentralization with Butler’s emphasis on shared precarity.Decentralized assembly further offers an alternative to the role of the individual as the structuring logic of liberal governance. Here the individual, neoliberal subject is counterposed to society and social modes of accounting for difference and inequality. As Wendy Brown argues, “If there is no such thing as society, but only individuals and families oriented by markets and morals, then there is no such thing as social power generating hierarchies, exclusion, and violence, let alone subjectivity at the sites of class, gender, or race.”11 For Brown, neoliberalism’s intensified emphasis on the individual as the locus of both responsibility and self-determination undermines the potential of collective equality and shared responsibility at the heart of the democratic model, an ideal that for her has never materialized in its complete form. Nonetheless, a framework in which “we’re all leaders” is one in which the entire assembly is simultaneously responsible for the presence of “hierarchies, exclusion, and violence” and empowered to change it.It should be noted that all of the above qualities have been powerfully expressed by the various assemblies themselves. For example, Ferguson Action, the coalition group that arose from the protests of the murder of Michael Brown, described themselves this way: “We are connected online and in the streets. We are decentralized, but coordinated. Most importantly, we are organized.…We stand beside each other, not in front of one another.…We can’t breathe. And we won’t stop until Freedom.”12 Moreover, all of these qualities—physically present although connected with media, radically horizontal through an emphasis on the group over the individual, manifesting performative politics via bodies on the street—are to some degree interrelated. It is also worth pointing out that the above features common to popular assembly over the past decade are not historically unique, and are in many ways essential characteristics of assembly more broadly.13 If anything, the past decade stands out more for the number of different movements that have claimed the rights to assembly rather than any particular novel quality that these movements bring to the table. Even the relationship between these movements and the various forms of digital media they use has clear precedent in other historical moments and older forms of media.14The connections between political assembly and media take on a number of different valences, from media produced by assemblies to media produced about them by mainstream journalists and others, to forms of media that exist in some symbiotic relationship with the political project of the assembly itself. This last valence is productively explored by Kyle Parry in his discussion of “generative assembly,” a formulation that draws on many of the threads discussed above to analyze the role that various media collections play vis-à-vis political conflict and social protest. Through a close reading of different archives and collections tracing the Hurricane Katrina disaster, Parry demonstrates that the process of collecting, compiling, and ultimately assembling media artifacts is an “especially powerful means of intervening in prevailing conditions of representation and remembrance around events of environmental and social violence.”15 In other words, the assembly of objects can perform part of the political work undertaken by an assembly of people when placed within this discursive context. My contention here is that the boundless photo assemblies that appear on Google Street View and elsewhere belong in the same political register as human assemblies.16The photographic image has long been theorized as an extension of pictorial perspective, a mechanical automation of the camera obscura’s contribution to Western visual culture via the quattrocento’s reproduction of space—and further, a manifestation of the larger ethos of Cartesian rationality and commodity exchange at work in the culture that produced it.17 In the 1970s, film scholars under the heading of apparatus theory extended this lineage to include the film camera and the ideological, subjectifying potential of moving photographic images to shape the spectator and their psychic world view.18 Later, scholars such as Anne Friedberg working on early digital media would extend this connection to further include the visual regime of the computer interface.19To be sure, the received narrative connecting the camera to the camera obscura and photographic perspective to the self-sufficient Cartesian subject remains contentious. Scholars like Martin Jay and Svetlana Alpers have challenged the singular hegemony of Cartesian perspectivalism throughout its history by demonstrating the existence of alternative visual regimes.20 Others, including Jonathan Crary and W. J. T. Mitchell, have argued that radical breaks divide this smooth continuity, separating the computer from the quattrocento or the modern eye from its Renaissance roots. Nonetheless, the connection between photography and perspective has largely persisted, and forms the terms around which we engage the image into the digital age.21 Susan Sontag and Ariella Azoulay, for example, may differ on the question of whose perspective the image captures or conveys, but they both maintain that the individual perspective (of the photographer, the photographed subject, the viewer, the lens) is the currency of the image.22Given this, it is perhaps unsurprising that the emergence of tools and methods for combining multiple images and exploding the boundaries of the frame frequently conceive of this move in terms of perspective. A telling example here is the series of photo assemblies that the artist David Hockney created throughout the 1980s. After primarily painting for nearly twenty years, in 1982 Hockney began a series of multi-photographic collages that he called “joiners.” Initially working with a Polaroid and laying out the images in a standardized grid, he experimented with a variety of different capture and assembly methods over the next several years, with the process eventually culminating in 1986 in his iconic image Pearblossom Hwy., 11–18th April 1986, #2.23 Hockney’s interest in these experiments was about the camera’s ability to “see geometrically” and in finding a way to use that connection with perspective to capture multiple conflicting perspectives in a single (assembled) image.24 As computer scientists and others began creating algorithms capable of stitching together multiple images, Hockney’s joiners provided both an inspiration and an illustration of the challenges computer vision would need to overcome in order to correct for perspectival differences and efface the seams between images that Hockney worked so hard to include.25As digitally assembled image combinations (both still and moving, real and virtual) have made their way into a variety of hardware and software interfaces, they bring with them the preoccupation with perspective and the connection to older forms of media it entails. Oliver Grau and Erkki Huhtamo, for example, place Virtual Reality (VR) and other immersive (assembled) image technologies into a much longer history of architectural spaces and panoramas that utilize multiple images to reinforce the illusion of continuous space, often with very specific political intent.26 Here, linear perspective evolves to fit curved canvases, screens, and other surfaces, thereby rendering a type of multi-perspectival view, although one where the individualized spectator remains the object of address.Throughout the 1990s and early 2000s, digital algorithms made strides toward the automatic assembly of multiple photographic images, developing specific procedures to hide the resulting seams and overcome differences in exposure, camera angle, and the conflicting perspectives that defined Hockney’s joiners. Eventually, these algorithms made their way into various consumer devices and platforms. Around 2000, Nikon, Canon, and other manufacturers began introducing shooting modes on digital cameras capable of stitching together images in-camera or later via branded software on a computer. In 2003, Adobe’s popular image editing program Photoshop introduced its Photomerge tool, which was designed to simplify and partially automate the production of assembled images. In 2006, Microsoft previewed the Photosynth platform, which was capable of stitching together up to six hundred images into an apparently seamless, navigable space, and just a year later Google launched Street View, previewing what would eventually become the most boundless photo assembly to date.As these image assemblies proliferate, the connection between the individual image and perspective begins to shift. For William Uricchio, Photosynth and similar platforms signify a decisive break because they destabilize the fixed and privileged viewing position of the single image, introducing a multi-perspectival space that upsets the subject/object relation that had defined images up to this point.27 For Uricchio, this same algorithmic turn was reflected in other digital creations like Wikipedia that privilege the group over the individual. In a similar argument derived from a radically different domain, Shane Denson has asserted that the appearance of what he refers to as “crazy cameras” in the aesthetic space of cinema upends the traditional relationship between spectator and apparatus.28 By transferring between such views as traditional cameras, security cameras, and a wider array of non-optical imaging devices, these cameras and the fluid, malleable images they produce challenge the fixed perspective (and hence, fixed subject position) of the cinematic image.The emergence of the “dis-correlated” image in narrative cinema (or “post-cinema,” as Denson would have it) highlights an important facet of the image assemblies that both Photosynth and Google Street View produce: namely, the intrinsic fluidity of the image and its integral connection with a particular viewing platform. As Jihoon Kim convincingly argues, how we interpret the subject/object relationship at work in something like Photosynth depends on how and where these images are viewed.29 Moving from the immersive repose of a VR headset to active navigation with a smartphone on a busy street corner produces a radically different image as well as a constitutively different relationship with the subject who views it.30 Like the upscaling smart television that Denson describes, these interface devices alter not just our perceptual experience of a fixed image but play an integral part in its algorithmic (re)assembly.Both human assemblies and photographic assemblies arise out of a diverse set of discrete entities and foreground the group over the individual, often effacing particular elements of any given person or photo in the name of larger cohesion. And just as human assemblies eschew traditional leadership, so assembly platforms like Photosynth offer “no correct or authorised viewing position, no ‘master shot’ within which everything else is a recomposition.”31 And finally, both human assemblies and photographic assemblies evolve over time, changing as events on the ground or the constituent parts of the assembly shift.Beyond the various procedural and material resemblances already elaborated, human assemblies and photographic assemblies collide and collude in other ways. Indeed, many of the theorists cited thus far have explored either the political stakes of various media forms or the mediated nature (in the strict sense) of human assemblies. Positioning photo assemblies as an extension of or break from the longer arc of Cartesian perspectivalism and its assumption of the unified subject already positions these images within some sort of political nexus. And no mention of the human assemblies that have emerged in the past decade would be complete without some articulation of the relationship that they form with smartphones, social media, and mainstream media coverage. But what I want to explore here is the way in which these sets of politics, the rules that govern photo assembly and the legal claims of specific human assemblies, directly work to reinforce and rearticulate one another in unexpected ways.We can see this interaction at work, I contend, in Google Street View and its “coverage” of two wider human assemblies from the past decade: the housing crisis that has emerged across the US and the Black Lives Matter movement and its recent occupation of space in Minneapolis, Minnesota. Alongside its Earth and Maps projects, Street View comprises a massive dataset of images integrated with numerous server and client-side protocols to produce high-resolution, 360-degree panoramic imagery from around the world.32 To capture this imagery, a number of different custom-built, car-mounted camera rigs were developed and deployed to drive around city streets collecting images and vacuuming up data. The current iteration combines images from fifteen cameras with GPS and lidar data to construct a series of largely seamless photo assemblies.33 The effect is of a single, seamless photo that one can visually navigate in multiple dimensions (both spatial and temporal).Street View poses significant privacy concerns. These run the gamut from embarrassing or illegal behavior inadvertently captured as the cameras rolls past to the unexpected implications of creating a massive new dataset that quantifies various characteristics about neighborhoods, cities, and the communities that inhabit them.34 Moreover, it bears repeating that all of Google’s navigational applications (Street View, Earth, and Maps), whatever their merits, are built to harvest data from both the images they capture and the users who view them in order to target and sell advertisements according to the extractive logic that Shoshanna Zuboff calls “surveillance capitalism.”35 While Google deserves credit for making many of Street View’s image assembling algorithms open source, its APIs also feed data back to Google, making it the epicenter of the data flows that Jason Farman calls the “digital empire.”36 The irony here is that Google willingly blurs specific features like faces and license plates to protect the privacy of the people who appear in the images it captures, but readily tracks these same characteristics (identity and movement) in virtual space through its ownership of highly valuable “free” tools that make its online properties among the most trafficked in the world.This model of proprietary data extraction from public space and user surveillance is thus part of the politics immanent in the algorithms that capture, construct, and display the massive, boundless photo assemblies that provide the core of Google Street View. This politics finds a direct corollary in the common integration of Street View images into real estate websites like Zillow and Trulia, a space where physical and virtual forms of private property and capital converge. Indeed, Street View is very much defined by the description and navigation of public and private space, using the photo assemblies it captures and deploys as the transit point between data property and physical property.But while this is the dominant political framework that governs the photo assemblies that define Street View, the street itself has always been a radically contested space, one that challenges established definitions and frameworks. As Butler puts it, the street is where we establish “what it means to assemble in a crowd…and what it means to move through public space in a way that contests the distinction between public and private.”37 As Butler’s point makes explicit here, it is human assemblies that enact and perform this alternative political framework.We can see the conflict or disjunct at work if we consider the various ways human assemblies manage to disrupt the smooth, one-to-one correspondence that Street View aims for by altering some facet of the spatial, political status quo, or what Ali Aslam has described in terms of the aesthetics of the ordinary.38 Indeed, one of the primary tools human assemblies deploy is the occupation of public spaces like streets, squares, freeways, intersections, and other spaces of transit and movement—i.e., the very activities that Street View is meant to enable. Of course, these events don’t make it into Street View because they’re typically too ephemeral to be captured by the semi-occasional rounds of the Street View cars. Not intended to be a real-time video feed, Street View is notoriously bad at dealing with movement more generally. Many of its impressively infrequent stitching errors result from moving objects, an affliction of algorithmic photo assembly more broadly.39 In this sense, its images are antithetical to movements generally, a tool of the status quo.Consider, for example, the massive, worldwide wave of Black Lives Matter protests that erupted in May of 2020 after the murders of Ahmaud Arbery, Breonna Taylor, George Floyd, and others.40 In spite of the constraints of COVID-19 stay-at-home orders and social distancing guidelines, millions of people assembled in person to register their outrage. In Minneapolis, at the intersection of Chicago Avenue and 38th Street, the site of Floyd’s murder, the assembled group occupied several square blocks. Over the next several weeks, the space slowly transformed as memorials and art projects appeared in honor of Floyd. Over many months, protestors held this space through their presence, even erecting and staffing makeshift outposts at entrances to the intersection to keep people warm during winter.41 As of this writing, the intersection, now known as George Floyd Square, remains closed to vehicles as protestors and city officials negotiate how to preserve the space as a more permanent public memorial. While it is far from solving the long-standing problem of racialized police violence, the transformation of space at the demands of protestors represents a significant achievement, and a prominent reflection of the political power assembled groups wield.And yet, nearly a year after it emerged, Street View still displays the space as it appeared before Floyd’s murder: a generally nondescript intersection flanked by a gas station and various retail stores. While there are many obvious, plausible reasons for this (the infrequency of updates to different locations, the impossibility of driving a car through the space while it was occupied, etc.), the disconnect between the work of the human assembly and the record of the photo assembly reveals the conservative nature of Street View and the limits of the world that it claims to allow people to explore.A slightly different disconnect between the politics of Street View’s boundless assembled images and human assemblies appears if we consider its largely inadvertent documentation of the increasingly ubiquitous ad hoc settlements that now define many modern urban spaces. While these encampments are largely inhabited and occupied out of sheer necessity, many end up doing the sort of political work that Michelle Lancione refers to as “‘radical housing’ within everyday practices of dwelling at the margins.”42 Given the economic fall out of the subprime mortgage crisis in 2008 and the sharp rise in evictions that Matthew Desmond and others document, as well as the ongoing economic inequality that characterizes urban life under contemporary capitalism, housing now figures as a primary point of political resistance.43 Encampments, alongside other precarious arrangements such as vehicles and the “illegal” occupancy of vacant property, figure as part of what Ananya Roy describes as a form of “strategic illegality—as a politics of emplacement.”44 In this framework, the occupancy of a particular space expresses a political right to exist in the face of larger ongoing forces of displacement and disenfranchisement. Homeless encampments and other forms of radical housing do the work of political assembly, asserting a claim to legitimacy and belonging within urban communities.In many ways, they also confound the drive toward classification and property relations that propels Street View. For example, consider the Street View of the off-ramp that circles Interstate 80 at the University Avenue exit in Berkeley, California (Exit 11). Interstate 80 provides one of the major routes leading into and out of San Francisco, branching off into Oakland and, further on, San Jose, regularly ranking among the busiest freeways in the Bay Area megaregion. For many residents and visitors to Berkeley, Exit 11 works as a sort of unofficial gateway to the city, offering a main artery to Downtown Berkeley and the University, while connecting the city with the Marina and waterfront trails via pedestrian and vehicle overpasses. The ramps here form a cloverleaf pattern, leaving large swaths of land on either side of them as they circuitously route drivers onto and off of the freeway in both directions.As the housing crisis in California and elsewhere escalates and ongoing economic disparities push greater and greater numbers of people into unstable and makeshift housing arrangements, the area around Exit 11 is the site of one of the many encampments that now define American cities. The encampment at Exit 11 is one of fifteen across Berkeley that, as of July of 2019, housed more than one thousand people, a 34 percent increase in just four years.45 Because of its highly visible location and several dramatic events that have attracted news coverage, the Exit 11 encampment has come to symbolize the larger problem around homelessness, poverty, and the inability of the city to easily resolve the ongoing and long-standing structural disparities that fuel it (including massive income inequality and capital exploitation). Over time is has become known as the “Where do we go?” community because of a local activist group that emerged from its members.46 While the area is geographically connected with the city, the land is technically owned by Caltrans, the statewide transportation agency responsible for freeways, highways, and public transit. Because of this, Exit 11 falls into a sort of administrative liminal space, with both the city and Caltrans disclaiming responsibility for both the people and possessions currently there.47While impermanent dwellings such as the “Where do we go?” community should theoretically suffer the same fate as George Floyd Square in the Street View universe (they are frequently cleared by police and other city officials), their sheer ubiquity contingently lodges them in the camera’s view. Street View has captured various iterations of Exit 11 over the past several years, and comparing them using its “historical views” feature reveals something of the paradoxically ephemeral permanence that characterizes them. Various tents appear and disappear alongside possessions, donations, and the trash that’s illegally dumped there by non-residents. Encampments like “Where do we go?” along with the other forms of radical housing discussed above defy the logic of Street View by connecting public space with private life, vehicles with dwellings, forcing us to confront, either in person or on the screen, the boundaries of communities.By way of a conclusion, I would like to turn, if only briefly, to an alternative mode of photo assembly that manifests a radically different aesthetics of assembly. Since 2004 (roughly the same moment when Street View and Photosynth appeared) the Japanese photographer Sohei Nishino has been capturing and composing large-scale photo assemblies of various cities in a series he calls Diorama Maps. His method for each diorama is to spend several weeks or months walking a given city and photographing whatever draws his attention, recording several thousand images in the process. These are later developed, printed (he still uses film), and laid out in a rough representation of the geographic space of each city. Similar to Hockney’s joiners, Nishino’s assemblies are manually constructed and foreground the various seams that connect their constituent parts. Juxtaposing street-level images and individual details with wider landscapes and vistas, the final impression offers a glimpse of what might emerge from a database glitch that stitched Google’s Street View and Earth projects together. While each individual image obeys principles around perspective, the larger diorama reflects shifts in time and space as one travels the terrain of urban geography. For Nishino, this reflects the appearance of the city as filtered through his own experiences and impressions. This is “not at all a map to convey precise information, but the record of how I, as a human being, have walked through their streets and how I looked at those streets.”48I invoke Nishino’s photo dioramas not to erect a simple binary that valorizes artistic output over commercial products or analog imagery over digital imagery, but rather because these images so powerfully foreground what Street View seeks to conceal: evidence of their construction and the mediation at the heart of photo assembly. Street View works very hard to present a boundless, transparent representation of public space, and the largely seamless, mundane view it offers is an astounding technological achievement. And yet, this representation inadvertently reinscribes the dominant political frameworks that govern the physical spaces it documents, areas where public and private, citizen and state, boundary and border all collide. As human assemblies increasingly take to the street to challenge this political framework, they inevitably alter the seemingly stable landscape that Street View documents, a reality its photo assemblies will sooner or later have to reflect.