Image from, adopted by horse_ebooks

Selfhood, the icon, and byzantine presence on the Internet

Everest Pipkin
16 min readMar 12, 2016

This is a considerably-expanded version of the talk I gave at Botsummit 2014. The video of that talk is viewable here.

Very Long Ago

When we talk about technology, it is so often in the rooted in the Short Now, but I would ask you to cast your thoughts backwards. I’d like to begin in the Byzantine Empire, 11th century. Emperor Basil the second has just died, and with his eclipse begins the slow imperial decline. The Great Schism, the theological split between Church in Rome and the Church in Constantinople, is looming on the horizon. Soon, the crusades will shake the structure of regional power and daily life to the core.

When speaking of this time, it is important to hold in mind some fundamental qualities of the Byzantine worldview. As starting point, one can consider Patristic thought, which was the philosophical basis for much of the writing of the time. Patristics is a best-of-many-worlds situation, and owes its roots to first-century Rome, taking lessons from the Old (and, later, New) Testaments, the apostolic teachings, Greek philosophy and Gnosticism. This many-historied tradition is important to us in the 21st century, because it helps allow us to unlock this ancient (and therefore quite foreign) worldview through other philosophies that have come to bear more heavily on Western contemporary culture than that of Byzantium.

Byzantine philosophy always referred to what is beyond experience and nature, to the existence of God and to the real being. Byzantine worldview has shifted the Platonic distinction of the Greeks between the intelligible and sensible world (the difference between the ‘here’, this wooden block I hold in my hand; and the abstracted platonic ideal, the ‘there’, the idea of the wooden block). This split instead points to the distinction between the created and the uncreated being. The created (human civilization, most works of art, economics, the natural world, etc) can perhaps point to the uncreated world, but is fundamentally tainted by its mortal origin and may never reach divinity. In contrast, the uncreated is intrinsically all that is divine. These uncreated energies include the presence of God in scripture, the human soul (without a life yet lived), and various miraculously created artifacts and icons that received particular veneration in the church, called acheiropoieta (which were also known to replicate themselves). The ‘uncreated’ is a tricky subject for contemporary readers, as there is little analog today. It is important to remember that the uncreated and created lived together on Earth, even if their constant connection was generally invisible. It was in fact this tension and desire to grasp the invisible, either by discursive reason, artwork, or by faith, characterizes Byzantine metaphysics.

In short; this is a space of intersectional worlds — divine and mortal are on one plane, touching always.

Perhaps the clearest example of this is illustrated in one of the great arts of the Byzantine Empire, the icon painting. After having looked at several icons of the era, one may notice that these objects only seem to exist on a flat plane; all depth is absent. This is not a mere aesthetic choice- the “third” dimension of an icon painting is not space, but spirit. The realm of the immortal intersects this one, the mortal one, at every point in space and stretches to an eternity; such signifiers have no need for an artificial, painted depth. They are not windows into the divine but literal divine objects that exist everywhere, but are viewable in this ‘thin space’- the thin space that an icon gathers around itself, cutting through.

Iconographic paintings are not signifiers, but are physical structures that exist in both worlds; a picture of a saint is not a picture of a saint, it is the saint, physically in space. The uncreated (the divinity of a saint, of God) is there, in the icon, along with the created (the painting by human hand). They are tools to grasp the invisible.

Much Later

Great, okay, so, here we are a near perfect millennium later halfway across the world, considering the the act of painting god.

You’re probably wondering what this has to do with the internet.

I’d like to touch base on some more recent history, the perhaps-obvious choice of Walter Benjamin’s iconic 1935 essay on a mechanized field of aesthetics, Art in the Age of Mechanical Reproduction. Benjamin claims that in the past, the role of artistic production has been to provide a magical foundation for the cult (or the spiritual). He claims that, then, value was located in its central position within ritual and religious tradition. A statue or idol has a detached authority and power, which is implicit only as rooted in time. The mass reproduction of such an object was not just unlikely; it was unimaginable. To remove it from its history was to render it useless to itself. It would be to remove the aura, that detached and transcendental power. This aura seems to rest on something autonomous, and the statue is not like any other object produced or used within a society; it appears free from ideological control or human interference, as though its power issues independently from within.

He goes on to say “mechanical reproduction emancipates the work of art from its parasitical dependence on ritual.” Mechanically produced objects begin as almost-identical duplicates and are placed in the world as unique entities with individual existence. Their origin may be cheap (the factory), and as Benjamin argues, perhaps that does constrains their aura to the time that will accumulate on them and not their mere existence. They gain their individuality by living unique lives, but begin with little aura at all.

But digitization is not mechanization, and duplication within digital space is not autonomy.

A digital object appears across a multiplicity of screens both at once and forever. These entities are not individually manipulatable, which is to say they do not exist as individual duplicates that can accumulate a singular history and aura. Instead, digital objects exist singularly, everywhere.

For example, a website- lets say, the Wikipedia page on relics- may be accessed both by you, and me, and 150 other people uniquely stationed around the world, all at once. We are not having a shared experience. Which is to say; we are not made aware of one another, we are not in one room beholding an object. The page, however, has not duplicated. There are not suddenly 152 Wikipedia pages on relics that have separated out, formed unique identities for each of our screens. There is one page; and if any of us choose to edit the information available there, it would populate that change across all screens, and carry it forward, into the future; a single entity, accessible in multiplicity.

This is not to say that digital media may not be replicated. However, digital replication- the act of making a copy- does not work in the same way that physical copies do. When Benjamin talks about reduction of aura via copy, he exposes the possibility of cheapening. This is perhaps an awareness of the ‘non-specialness’ of objects, a knowledge that one’s object is not unique in the world and that aura is therefore built, not intrinsic. The cheap copy has become something of a trope, and the phrase can sometimes be used quite literally; the knockoff designer handbag or poached television premise both promise us the illusion of aura, at a discount. However, these copies have the remarkable capacity to reach backwards to their origin and damp that original aura and power. The real Chanel bag is suspect not in spite of but because of its many duplicates. The accessibility of that logo to anyone with $20 and an eBay account slowly mutates the Chanel product itself from unreachable to omnipresent, and therefore cheap.

But in digital space, a copy (a real copy, say an image accessed online but saved to a hard drive) is often absolute. It is a perfect recreation. Unlike the handbag, where an expensive (either literally expensive- in materials or labor- or conceptually expensive- branded) parent spawns many cheaper children, the photograph downloaded is generally exactly the same as the photograph online. There is no distinction between original and copy; they exist at the same scale, have the same aura. These copies instead function like mitosis; once separated, the child and the parent are generally indistinguishable. Digital objects bud, and neither is reduced for the reproductive budding. Even when the child downscales or glitches or is edited, the parent remains untouched by the change; here, a copy has the capacity for individual mutation but does not intrinsically affect its parent.

Information that is duplicated online (without, say the download or saving of information to a personal computer) has a similar structure. Although the internet is resplendent with copied (and sometimes stolen) information, images, and other digital objects, their originals are almost indiscernible from copies. They all exist at the same scale, on similar platforms, given moderately equal weight. In fact, stolen digital objects are unique from stolen physical objects in that the original object is rarely gone after the theft. For example, when a car is stolen, the owner of the car has an empty garage (a void where the car used to be) and a thief has a car. But when an image is stolen (perhaps a fraudulent individual claims a drawing is theirs that was lifted off an artist’s website), it is still very much on the artist’s website, as well as on a new Deviantart account, and perhaps a Tumblr without accreditation, and perhaps again as a header image on Facebook, and as a print-on-demand t-shirt, and so on.

But with tools like Google’s search by image, these are all connected, networked by their shared identity. When networked this way, such objects collapse back again into a single object. In fact, search engines are remarkably excellent at collapsing unique identity back into single idea. Copyright issues aside, a mass of duplicated images used indeterminately across a variety of websites are all one image. We see a picture and say “I have seen that before”. Text mirrored off of Wikipedia and onto any of a thousand Wikipedia-like informational sites is still Wikipedia text. Social media sites excel at this type of networked duplication. A retweet of information is not a shift in scale, or a real copying; a retweet impacts the structural bridge of a networked idea (furthering it outwards), but not the intrinsic idea itself.

Of course, there are endless exceptions to such rules, and many objects exist both digitally and physically at once, the divide being fuzzy at best. But the fundamental argument remains; although these digital entities are accessible from any place, by multiple people, at once- and even though these interactions may happen simultaneously- they are singular. Even when copied or stolen, such digital objects reference themselves, a cloud of idea that comes to form an individual Idea, the Idea of the image or the text or whatever else. Such entities escape the cheap copy, and exist both inside of and outside of the veneer of their accumulated time. In short; they have aura, the same kind of aura that a single object- unique in the world- can claim.

The shame of horse_ebooks (selfhood on social media)

The mechanical turk was a supposed-automaton (or robot) constructed in the late 18th century. The turk was carved of wood, dressed in mystic robes and seated at a low table. He was capable of playing (and winning) at chess, as well as completing the puzzle game ‘the knight’s tour’, in which a knight occupies every square of a chess board exactly once, visiting all 64 squares. The table and Turk contained many doors, in which an audience may see the complicated clockwork that drove the automaton. It was nothing short of a miracle. In the 1820s, the Turk was exposed as a hoax. Inside the table was a tiny hidden box in which a diminutive chess master may hide and control the man above. Destroyed by fire in 1854, the Mechanical Turk has re-entered contemporary lexicon with the launching of Amazon’s task-hiring website of the same name.

Horse_ebooks was an alleged spam twitter bot originally written to sell ebooks that gained mass popularity for its poetic and funny tweets. On September 24, 2013, it was announced that Horse_ebooks had become part of a multi-year performance art piece staged by Buzzfeed employee Jacob Bakkila. Bakkila had approached the original bot creator, Alexei Kuznetsov in 2011 with the intent of buying the account; Kouznetzov agreed, and since 2011, Horse_ebooks has been operated by Bakkila, who either hand-wrote or curated generated tweets. Before the revelation in September 2013, it had more than 200,000 followers. Horse_ebooks has re-entered the contemporary lexicon in the form of the _ebooks bot, a popular form of ‘vanity’ bot in which a twitter user inputs their own archive of tweets as fodder for autonomous recombinations of text with their own voice.

I wanted to touch on both of these stories because they are a useful way to start thinking about entity and selfhood without intrinsic humanity or person-ness. Both horse_ebooks and the turk were fascinating because they claimed to occupy a middle-space between ‘person’ and ‘object’; a theoretical bot so good that it transcended mere routine and gained a shine of real intelligence. Of course, that intelligence was a trick, and they are interesting also as exceptions- deceptions, really- from actual automata, which often has a shine of real stupidity.

I’m not sure why these two incursions feel traitorous- if anything, they should be impressive. Perhaps it is because they seem just possible, just on this side of miraculous. Perhaps it says something about how we engage with technology; happier to celebrate and lift up the things we can almost understand, than to dissect them into disappointment. But perhaps too there is an ease of granting selfhood to objects, to non-living entities that act just enough like living ones to warrant such treatment.

There are many such examples of personhood given to objects, most of which have little attempt at an outward, conversational intelligence. Think of the children’s stuffed toy, or the painting on the mantel, or a particularly important memento or keepsake; we instill each with its own person-ness. This selfhood is especially strong in religious objects; it is certainly present in the icon paintings previously mentioned, but also in practically any holy representation of god or divinity or power across virtually every spiritual culture worldwide (see; crosses, fetishes, sculpture, blessed food or water, gates or doors, relics, architecture, holy symbols, so on). I want to point out that I am not particularly religious and have certainly not studied religion with enough care to make trustworthy sweeping generalizations. But these are certainly objects that have agency. They can heal, harm, protect, perform miracles, reproduce themselves. It would be a remarkable thing to call an object with such power ‘dumb’.

Horse_ebooks and the mechanical turk took advantage of our penchant of selfhood-giving, but both also rose on platforms that supported their premise. These platforms allowed the idea that each was capable of existing at all. The mechanical turk was disguised as an automaton, a robot that futurists everywhere have long dreamed of finally granting real intelligence. And automata in the 18th century were remarkable: there was a screaming man being devoured by a tiger; an eternally praying monk; a duck that ate barley and defecated compressed grass some seconds later; tea-servants and other household puppets; a whole ensemble of various musicians playing various instruments at various courts. Chess was hardly a leap. Horse_ebooks also came up in a friendly environment, that of early-days twitter, which was already host to multitudes of automated content, from spambots to artistic projects.

But most importantly, both physical automata and social media profiles lend personhood intrinsically. The robot looks like a living thing, acts like a living thing, and is treated… well, a little bit like a living thing. Similarly, social media bots exist at the exact same scale as humans do on those platforms. A bot on Facebook or Twitter or Instagram has a profile picture, a timeline, a bio, and a history that is comparable to any human person using the service. They may not be anything more than automated process, but they are given a deference far beyond that of a program on a loop.

I hope it is clear that by choosing a pair of examples which are fundamentally fakes, I am not trying to expose any great conspiracy of humans-posing-as-machines. It is 2016 and we already have real social media bots and real robots. Programming a computer to play chess is practically children’s play, and there have been many (even successful!) attempts at automating humor. Instead, the appeal of such stories is that they point to a successful middle-space, a sort of cyborg mentality of the almost-possible, of the too-good-to-be-true. Here, a false truth is accepted not because it is human, but because it is precisely not; it is a kind of miracle.

On mystery

I am not the greatest programmer. Although there are areas I have studied intensely, I am still self-taught and there are giant structural holes in my knowledge that I find little incentive to fill in. This means that when working with code, I end up in territory I don’t properly understand with semi-regularity. It is not generally a big problem; bits and pieces fall into place, I flounder my way through a few off-kilter solutions, and when things begin to work, I don’t touch them anymore. Fine.

A benefit of not knowing how to do things right is that doing things wrong can result in unique solutions, which although often inefficient and messy, do produce unique results. Of course, there are plenty of times when this is not desirable (5 added to 3 should probably result in an 8), but few bugs are entirely without their benefit. It is rare that I end up with a finished project that does not include within it a stabilized or embraced error. There are also times when things break quietly. Not all errors are catastrophic, and these quiet moments of failure slip through the cracks of testing, to be called done.

There is beauty in failure. The sudden barring of internal process or broken function shows us something deeper about the pattern behind the designed process; in some ways, it can feel more true. We are given access to the internal mechanisms of the automata; we see the edges of the world. Rather than breaking the illusion of a perfect, designed process, these breakages feel autonomous, undesigned. Pulled undefineds, nulls and errs; grammatical problems; corpus unreliability; the shifting landscape these things exist in; false geographical boundaries; failed physics and other emulation systems: each holds its own truth. They are the uncreated within the created, in the same way that the icon painting holds both inside of its borders. Unintended, they feel miraculous; non-designed, they are divine.

The icon is transparent

Surprise as tool of the miraculous is certainly nothing new. Perhaps we can remember the acheiropoieta, the particularly divine byzantine artifacts known to reproduce themselves (or their contemporary analogy, the remarkable appearance of the face of Jesus on whatever food item, which seems to make the news cycle weekly). What is somewhat new is how we have engineered a world so primed to produce these situations; the technology of the internet and programmed space is ideal for such moments.

Thinking about icon paintings from a contemporary standpoint, they also feel remarkably technological. As tools, they act as screens; access points to a world that lives everywhere and also nowhere, or at least- nowhere physical, not a single point. This space was deemed divine because it was constant, accessible, and a thing-into-itself; not a signifier for a not-present idea or a person, but actually present. The divine suffuses all things, and we can cut through normal space to expose it, access it, be transported through it. The icon painting does not require the illusion of painted depth because that depth is intrinsic; that depth is the entire invisible world, suddenly held.

The screens with which we access the contemporary content of the internet are also thin spaces. The world in which we live is laid over with another space that is accessible from any point where there is cell phone reception, wireless internet, satellites. We hold a screen, and see through it to another world, that of information. Screens are flat, but their depth is intrinsic. When we talk about being online, our language is geographic and physical: we say we are surfing, digging, exploring.

Many of us (those that live in cities, in connected countries, with the financial stability that guarantees an internet connection) live here, in this other space, as much as we live in our bodies. We say that we are on Twitter, on Facebook. We build digital selves that exist on these platforms whether or not we are currently accessing them. Because we grant our own power of selfhood to these systems, the beings that live solely in the digital world also adopt that power. The bots and processes that use social media or other human-intended services act as beings-unto-themselves.

That selfhood is perhaps not unlike the selfhood of the icon painting. There, the image of a long-dead saint acts as portal to a realm where they are very much still living, very much still empowered. Because the rules of that invisible world behind the screen of the painting are not quite the rules of the mortal realm, the saints live on, granting miracles, speaking to God. But as the invisible, immortal world connects to this one at every point, the miracles are physical. God acts with potency, not in the world of spirit but that of flesh.

The beings of our internet also reach out. They interact with us, leave comments or send us spam emails. They also steal our passwords and credit card information, ransom our files. They populate print-on-demand services with products, physical objects that can be bought and will arrive at human doorsteps. Sometimes, they both are and also make artwork. They are digital information with auras. They produce. They exist.

There are many frameworks with which to view the internet and internet-based entities. It is a network, or a rhizome, or a spiderweb, or a landscape. Information layers or gathers; people surf or browse. It is a repository of information made by linkages. It is the salvation of contemporary humanity, or a branded, controlled, vehicle of consumption. Perhaps there is truth to all of these things.

When speaking of technology, it is easy to forget history, to feel that we are in a time entirely new. This is a common trope of technological discussions; we say things like ‘at the dawn of the information age’ or ‘for the first time ever’ or ‘society has fundamentally changed’. And we are not entirely wrong; there is much that is new. But humanity has been grasping at the invisible for a very long time. We give power to our objects, and they become windows to the divine. We have sought the autonomous and the miraculous, felt the pull of holy places and gasped at relics and other things of beauty and power that teeter between the uncreated and the created. These are frameworks that are embedded in our histories, our language, and our understanding of the world. It is perhaps no surprise that we have built a system which intrinsically mirrors this long history. We have made a network that lays over our tangible world; we hold screens that cut through physical space to this other dimension; we live side-by-side with beings without body and feel their power. We made tools to hold the invisible in our hands.