Early in my career in search engine optimization, I was working with an SEO consultant on a particular website project, who gave me some words of wisdom: A website is a building. A domain was brick and mortar, a sitemap was a hallway, webpages were a series of rooms.
It was a compelling metaphor, one I kept returning to as I worked on optimizing and organizing the site’s pages. But the project became more complicated the longer I worked on it: Instead of finding a neatly ordered site structure, I found chaos. In the CMS were hundreds of broken links. On the blogs, old branding sat uncomfortably next to new design. Orphaned pages were hidden and sealed off in various, unexpected locations. There was, in other words, a monstrous version of the site lurking just beneath the public-facing one.
Encountering this hidden version produced the same uneasy feeling as looking at “horrible architecture,” a term developed by Singapore-based architects Joshua Comaroff and Ong Ker-Shing: structures in which “normal anatomy grows deviant — extra limbs appear, holes open where they should not.” In Horror in Architecture, their 2013 work of architectural theory, they use the conventions of horror cinema to read buildings, applying horror genre tropes — like the partially dead (the figure of the zombie, for example, or the haunted house), or the double (the doppelgänger) — to interrogate different architectural forms, from stave churches in Norway to the Twin Towers in New York City. It’s a bizarre and enchanting project designed to change the way we think about our built environment, to draw our attention to what we might ordinarily look past.
Horrible architecture, according to Comaroff and Ker-Shing, has multiple defining features, including “doubling, reiteration, disproportion, formlessness, shifts of scale, excess parts and openings.” Throughout the book, we encounter architecture that deviates from common design conventions. We expect residential buildings to contain windows we can open, for instance, but their examples include houses with false windows; mid-century apartment blocks with idiosyncratic modern extensions; hotels with sealed-off rooms and hidden voids. Here, what we see is not what we expect to find; these buildings embody “the suspicion that official design discourses are not telling the whole story.” Comaroff and Ker-Shing’s readings are underpinned by this suspicion: What is hidden from us in the buildings we occupy? And what does deviance reveal?
When I read their book, I thought first of the monstrous website I had worked on years earlier. Applying their suspicious gaze to the internet, familiar sites begin to suggest something sinister. Horrible architecture might best be read as an aesthetic framework for interrogating all contemporary structures; for identifying what Comaroff and Ker-Shing call the “creeping unease” at the center of modern design, namely the anxious intolerance of deviance. By looking at this deviance directly, we become more alert to the mechanisms used to conceal it; and to the internet’s fallibility, disposability, and ultimate mortality.
Horror in Architecture describes nine typologies of “the horrible.” The chapter titles read like a list of strange, dark poems: “Doubles and Clones”; “Exquisite Corpse”; “Partially and Mostly Dead”; “Reiteration and Reflexivity”; “Incontinent Object”; “Trojan Horse”; “Homunculism and Gigantism”; “Solidity, Mass, Stereotomy”; “Distortion and Disproportion.”
Early in the book, Comaroff and Ker-Shing write that “the horrible” arises “wherever there appears to be a perceived erosion of the conventions and structures that hold the seams together.” In other words, horror emerges whenever there is a rupture of some kind. Each of the nine typologies represents one such rupture: “Exquisite Corpse” refers to a break in convention, where the mixing of incompatible elements creates a monstrous whole (think of Miami’s Museum Garage, which combines the radically different styles of five different architects). The “Trojan Horse” is a rupture between the exterior of a building and its interior, where a “creepy, sometimes malevolent” internal space is kept hidden (such as the hidden bunker in the basement of the Greenbrier Hotel in West Virginia).
Horrible architecture might best be read as an aesthetic framework for interrogating all contemporary structures
Of all the typologies, the “Partially and Mostly Dead” is perhaps most familiar from the horror genre. The vampire or the zombie is neither fully dead nor fully alive; the haunted house is a space in which life and death coexist. Each represents a “mixture of life and un-life in the same object.” In architecture, this “unnatural” mixture finds expression in buildings that display signs of decay, ruin, and dereliction; spaces which are neither fully occupied nor fully vacant.
Comaroff and Ker-Shing distinguish between the “mostly dead” and the “partially dead.” The former, they argue, is the horror of the non-inhabited structure: Take the abandoned shopping mall with vacant shops (they cite the Sherman Oaks Galleria in Los Angeles), or empty houses in villages transformed into ghost towns by military conflict (like La Unión Peneya in Colombia). These buildings contain few, if any, signs of life. The partially dead, on the other hand, are the structures that are partly inhabited, partly abandoned: second and third-floor rooms of high-street buildings, for instance, are often left empty in German cities like Bonn. In hotel renovations, “decanting” refers to cutting away the floor of guest rooms in favor of more profitable spaces like restaurants, and inserting a ceiling board in their place, creating what Comaroff and Ker-Shing call a kind of “ghost room” or sealed void. The partially dead building, then, is a structure that is both “occupied and vacated at the same time.” A dead part of a live whole.
Online, we regularly encounter websites that display signs of decay, ruin, and dereliction. The 404 error — “page not found” — is maybe the most common example of the “partially dead,” being the “dead” part of a live site. This is horrible architecture: evidence of faulty construction, a rupture in our expectation of smooth navigation; the URL does not behave as it should. The 404 error is a fundamental part of our online experience, yet it never fails to surprise, bewilder or frustrate.
Not only does this “dead” page exist on a live and otherwise functioning site, but signs of life often haunt it. Multiple websites have attempted to give their 404 pages a human touch: the New York Times’s 404 page takes a personal tone, encouraging users to search the site for something else instead (“We’re sorry, we seem to have lost this page, but we don’t want to lose you”). Twitter’s now-discontinued “fail whale” depicted a contented-looking beluga being lifted out of the sea by birds, looking playful and optimistic — a reminder that the page was created by real people with a sense of humor. Still, the page points to a lack of structural integrity that cheery copy cannot mask. A URL has been changed somehow, either edited without a redirect link, or its content deleted entirely and its internal links not correctly removed. It’s a sign of ruin in an otherwise living space.
The partially dead building is a structure that is both “occupied and vacated at the same time.” A dead part of a live whole
Of the 1.5 billion websites online today, fewer than 200 million are active. The internet is populated with “mostly dead” sites — domains that are accessible to the public, but vacated. When I visit teen magazine Rookie, once a favorite site of mine, I find a red banner reading “THIS IS AN ARCHIVE. THIS SITE IS NO LONGER BEING UPDATED.” Its footer, 2011–2018, reads like a gravestone. On the homepage of the Awl, another old favorite, the featured article is titled “The Awl, 2009–2018,” with the tag, “and now it’s dead.”
“There’s nothing more disgusting than a publication entering senescence,” Choire Sicha, one of the Awl’s founders, told the Verge at the time of the site’s demise. “Everything dies! Choose life!” I’m struck most of all by the language of revulsion, and the idea of the website itself as a sentient organism.
As Comaroff and Ker-Shing argue, the partially and mostly dead structure is typically a victim of capitalism: Abandoned shopping malls often contain independent businesses forced to close by competition from larger chains; dead “ghost rooms” are created by profit-driven hotel renovation schemes. The same is true online: Partially and mostly dead websites reflect an evolving digital media landscape, unsustainable profit models and declining advertising revenues. Rookie, as editor Tavi Gevinson wrote, closed because the site was no longer financially viable. For the Awl (and its sister site, cult favorite the Hairpin), advertising was the issue: “Scale has become increasingly important for securing large ad deals,” publisher Michael Macher said in 2018. “Less of those dollars are falling to indie publishers.” The fundraising efforts of bigger sites like Wikipedia (with its banner imploring users not to look away), or the Guardian (encouraging donations by reporting how many articles a user has clicked) read like mounting signs of decay.
The partially dead site is categorized by stasis — a lack of content updates — which are more obvious, of course, if the site is a publication. But other types of websites show other signs of abandonment. During the 2021 winter period of the Covid-19 pandemic, the NHS website’s copy remained unchanged for a month, before it was updated with new symptoms of the highly transmissible Omicron variant. These signs of ruin are evidence of wider neglect: the underfunding of public services, understaffing, an overall lack of sustainable resources. They register as especially sinister at a time when the privatization of the NHS — restricting free access to healthcare — is increasingly likely.
The architecture of the internet is rendered “horrible,” then, partially by the demands of capitalism: In its wake, we find signs of deterioration and ruin. Navigating a landscape of dead sites changes the way we look at living ones; clean, minimalist design only cloaks the evidence of inevitable decay.
A building contains different materials — wiring, pipes, ducts — hidden in its interior. When this material is exposed, its architecture becomes horrible. This is what Comaroff and Ker-Shing call the “incontinent object,” which “releases that which should be kept inside.” (They cite the Centre Pompidou in Paris, whose elevators, corridors and exterior tubing give the structure a “transgressive and scatological atmosphere.”) In horror, the incontinent object is closely tied to the body. As Julia Kristeva noted in her theory of the abject, we’re repulsed when what is usually kept hidden — blood, for example, or bone — is brought into view.
In architecture, interior designs — elevator or plumbing systems, for example — are a “mess” that often contradicts a building’s “resolved bourgeois” exterior. “In the modern era,” Comaroff and Ker-Shing write, “‘better’ architectures have generally been more successful in their attempt to hide this dark matter. But it has everywhere given rise to a crisis of concealment,” a frenzied attempt to hide what is aesthetically displeasing. This “crisis of concealment” is not only a concern of modern architectural design, but extends to the design priorities of contemporary fashion and interiors — an effort, as many have written, to reduce sensory stimulation in offline environments when online ones are so cluttered with information. Website design itself attempts to offset this content overload, stripping back what is unnecessary to make online navigation less visually overwhelming and content more digestible. This effort is driven by both contemporary aesthetics and evolving user experience practices.
The internet is populated with “mostly dead” sites — domains that are accessible but vacated. Attempts to conceal this chaos create a creeping sense of unease
Modern design online is increasingly minimalist, an effort to conceal extraneous detail. Content organization systems (blog tags or time stamps, for example) are frequently hidden from a user’s view; technical information (a heading or meta tag) is concealed in HTML code, and dense pages of text (like terms and conditions or privacy policies) are relegated to the end of a webpage on its footer menu. These structures are the equivalent of the plumbing system or cable network: tidied away, frantically concealed. We are accustomed to this minimalist, uncluttered design — the website’s surface cleansed of everything unnecessary — and jarred when the “revolting technical interior” is exposed; minimalist design only renders the contrast between interior and exterior more pronounced. This “mess” is familiar to web developers and SEO consultants, in the mismatch between the back-end and front-end of a website. A site’s HTML code or CMS is often a disorganized and incoherent structure, while its customer-facing front-end gives no sign of its disarray.
Exposure might look like pieces of a site’s code appearing where they are not supposed to — the coding of a link, for example, that is visible within the written content of a page — or a 3xx redirection error code. Unlike the 404 page — which is unique to each site, often treated as a content page, and specifically designed to be user-facing — this page is server-generated and standardized. It reveals the site’s technical structure to us — explicitly, its poor architecture — by making visible its brokenness. Hiding these errors is a matter of lowering user bounce rate, but the incentive is an aesthetic one too: an embarrassed scramble to hide visible signs of a site’s neglect.
Comaroff and Ker-Shing’s theory of the incontinent object reveals that our online environments are not separate from our physical ones; they are plagued by the same conditions of deviance and disrepair, shaped by the same conditions, and subject to the same failings. The “crisis of concealment” that Comaroff and Ker-Shing identify is not only the task of the architect, but of the developer: an attempt to cover what lurks beneath a site and mitigate the signs of disrepair. “Rather than creating the illusion of a transparent, well-working interface to information,” as Rosa Menkman writes, “the glitch captures the machine revealing itself.”
When the interior of a website is made visible, we are forced to reckon with its mortality; the possibility that poor maintenance will eventually kill the site itself. The exposed interiors are signs of impending death, and a reminder that the internet we take for granted is fallible and impermanent. The “crisis of concealment” manifests in the user through a demand for seamlessness, and a hyper-vigilance to disorder. Everything dies! Choose life!
For Comaroff and Ker-Shing, horror is “a desire to understand and value deviance; to be suspicious of fundamentals and appearances.” Horror, they write, is above all “a kind of tolerance, a willingness to look.” When we do, we become alert to what was already there, and what architects have scrambled to hide from view: chaos, uncertainty, and disorder. As the internet ages and evolves, this disorder is increasingly commonplace. Our online spaces are littered with ruins, dead websites, broken links, and the attempt to conceal this chaos creates a creeping sense of unease. What horrible architecture reveals, essentially, is the truth that we are constantly navigating a landscape of decay.
There are of course differences in the ways that buildings and websites degrade. Buildings require constant upkeep; when they are abandoned, they decay before our eyes. A decaying website, on the other hand, might look much the way it always did, until you notice the text declaring it an archive; or it might be obliviously “undead,” with obsolete design but the tone of a thriving entity. Abandoned buildings remain part of the landscape, but abandoned websites must be visited, or stumbled upon. They “lurk” apart from our everyday experience online. Applying the lessons of Horror in Architecture reminds us that they still exist; and that the sites we enjoy today will probably meet the same fate.