How My Drawings Destroyed Me

The rocky road to nurturing creative confidence and finding one’s voice

An outlet for creative and emotive expression. That’s what drawing once meant to me. It was my way of making sense of the world, of working through my emotions and finding meaning in the sometimes cruel, senseless acts of life — those of death and departure (the acts which hurt me most).

As a young boy, I would draw loved ones shortly after they passed. This tradition began with my Grandma, followed by my Grandpa and my beloved golden retriever. I remember feeling a sense of urgency while drawing, as if I needed to capture their likeness before it faded my mind. Once completed, I felt exhausted and emotional fatigued, as if I just relived their lives and deaths through the small paper window sitting before me. These early experiences with mark making reinforced the role of drawing in my life as an emotional act — one of introspection and self-sacrifice.

Drawing of my Grandfather after he passed in 2012. Nobody was perfect, especially not him. He was abandoned as a child and forced to live on his own. In time, he became an eye surgeon and practiced painting on the side. This art/science duality would eventually pass to my mother (an architect) and me. Growing up fatherless, I came to admire him as a self-made man despite his many flaws: a contemporary Miltonic hero. The harshness of the lines characterize well his forceful, mercurial personality.

It comes as no surprise then, that while studying art in school, I chose to further explore what drawing meant to me. For my thesis, I committed to a deep dive — a dissection — of the act of drawing itself, from both an artistic and a computational lens. It was only fitting, considering my parallel studies in psychology and computer science, and my ardent desire for exploring the intersection of art and science.

From the beginning, my goal was to use this project to better understand myself inside and out. Why does the pen under my fingers move the way it does? How much can I attribute the character of my marks to the physiology of my body? To what else might this character be ascribed? When all is explained away, what’s left? Will there be a soul remaining, lying prostrate on the paper?

A sample from an experiment analyzing drawing speed, pressure, mark styles and progressions through object hierarchies. In this experiment, I recorded drawings on a tablet, then afterwards generated colored graphs for analysis. The writing around the graphs are my annotations of observations. Specific experiments include (from left to right): (1) Graphing progression against a color spectrum; (2) Graphing mark length versus eccentricity; and (3) Graphing marks continuously where line width indicates pressure.
A synthesis of observations regarding the nature of cybernetic feedback loops in the act of representing visual imagery. This is one of hundreds of diagrams collectively defining the drawing model. The act of writing notes on large sheets of paper helped me collect, layout, and work through ideas. Every square in the above diagram indicates a map at a given time. I use the word “map” to define a 2D grid of activation / inhibition along some predefined spectrum. Maps often carried across multiple spatial frequencies to account for ranges of spatial complexity in image operations. Many of these diagrams focused on low and high level psychological models, and much inspiration was drawn from contemporary models of visual pathways and processes in the brain.
This diagram is one which focuses on the more physiological and physical models of hand, pen and paper interactions. It defines an approach for mark-making guided by goal-setting and accounts for some of the idiosyncrasies of my movements when switching marks mid-thought. In the end, the model defined marks as patches or lines, where each can be broken down into segments. Segments could be defined by a Goal, Guide and Gate that constitute the set of instructions passed to the motor system. The model dictates that cognitive modules must compete for limited resources. The primary modules consist of Apperception, Interpretation and Expression. Here, Expression is analogous to the motor system.

Quickly, the experiment obsessed me. I became engrossed in the extreme pursuit of self-understanding. I conducted countless tedious experiments on myself, trying to pick apart the idiosyncrasies of my body and mind. In time, I developed intricate, matchstick models of my inner workings. I could explain how my hand moved and why my mind moved it there as a complex physiological and psychological feedback loop. Combined with intent, this process converged on a satisfactory representation of my mental image of the outside world.

The model wasn’t perfect (not even close to it). However, this didn’t upset me. I came to appreciate the mind as a black box beyond certain levels of understanding. There were many decisions and representations I could feel present within me but existed just out of the reach of conscious comprehension. It felt like walking down a dark tunnel with a dim lantern: Eventually, I approached a thick fog that impeded my senses, preventing me from venturing further. The model made me realize how little I understand the motivation behind my behaviors, cloaked behind a veil of perceived lucidity and control. At once I came to realize both how deterministic my choices are and how little I truly understand myself.

A complete model of the algorithm and its input (imagery) and output (substrate) on the left and right, respectively. Circles indicate processes and squares indicate data. The color of lines indicates whether processes are reading from (blue) or writing to (green) data. The three primary cognitive modules are visible as groups of circles on the left, center and right, representing Apperception, Interpretation and Expression. Notice how the former runs once, followed by the latter two ping-ponging back and forth until the drawing is complete.

Looking back, I never stopped to wonder how the system I was observing might change through the act of observation itself. It never occurred to me until many months later, that I may have inadvertently changed myself, fundamentally and irreversibly. In essence, by assuming the dual roles of observer and observed, I couldn’t help but change the system — me.

As soon as the model represented a version of self, it became in that exact moment a version of past self. Slowly, I began to notice changes in my perspective of the world. In particular, drawing lost its special, intimate, sacred and magical place in my heart. The act of drawing became explainable: a logical function of enormous complexity, the likes of which I could not fully wrap my head around, but a function nonetheless.

When I tried drawing again, I could finally explain why I made every mark. However, this explanation bought me no joy; instead, it took all the joy away. I couldn’t stop the voice in my head calling out to me every moment, telling me why I chose to represent a line in this way or a patch of shading in that way, why I proceeded from one object to another, and so on. The magic of drawing receded away, and along with it, my interest in creative and emotive expression through the one outlet that meant something to me for so many years.

A time-lapse drawing made by the algorithm. The source of imagery was a photo of my face. The algorithm doesn’t draw this quickly; in reality, it took about 8 minutes to produce this image. Note the intentionality of mark type, line thickness, line direction, object progression and attentional window. The variation in these qualities reference the complexity of the model and motivations and choices guiding its stroke.
Another drawing made by this algorithm. This sketch also took about 8 minutes in total; this is a sped-up version. The source of imagery for this photo was my ear. The algorithm, like myself, finds second- and third-order curvatures, like those found in the contours of the ear, more interesting than rectilinear compositions. The algorithm is designed such that it only draws sufficiently interesting images; these two images passed the interest threshold.

Why was I so convinced of my behavioral determinism, when I simultaneously admitted an inability to fully understand myself? Why couldn’t I cherish the small hope that there is still a magic in the act of drawing? Because the experiments produced mountains of evidence in favor of determinism, which I couldn’t easily disregard. I live my life prioritizing hands-on learning, cherishing discoveries as self-evident truths. Disregarding my discoveries would mean going against my deepest beliefs. Naturally, I couldn’t turn away from them, so determinism became part of my psyche and my worldview.

The scientific half of me couldn’t stop ruining these once mystical acts, where from humble paper and ink, subdued to pain and persistence, emerged an image — a ghost — out of nothingness. This voice, emblematic of the pursuit of knowledge, felt like that annoying friend who explains movies while you’re watching them. But unlike that friend, the voice didn’t quiet down when I asked it to. Instead, it remained steadfast by my side, and with it came a profound distaste for drawing.

Somehow, I still found joy in others’ drawings. I could still marvel at them and the beautifully personal processes from which they emerged. Perhaps, I could still do this because others’ minds will always be a mystery to me. However, my own processes, once sacred, were now dissected and disassembled, shattered and scattered, laid bare on the table before me. Whatever aura of magic these bits and pieces one had, now faded away. From them, I could no longer derive pleasure, and I had no one to blame but myself.

I felt like a lost of part of me. A sadness grew within me, as I came to understand the error of my ways.

In the years that followed, I looked increasingly to technology as both a source of inspiration and a tool to mediate my creative endeavors. In particular, I became entranced by robotic arms. In them, I discovered a fascinating phenomenon in the unique way they moved. I began by asking them to draw spirals. Much to my surprise, repeatable artifacts emerged from their motions — artifacts that revealed the “character” of the machine. I saw my role as a humble servant to the robot and I worked to capture these fingerprints with as much precision and impartiality as possible. I sought to minimize my own biases in their creation in order to afford the robots an independence and the drawings an integrity.

These drawings were beautiful because they were magical: their source was a mystery to me. I could not explain them, so I reveled in them. I eventually developed some inkling as to the origin of their emergent patterns, but I stopped myself early on from trying to solve a mystery whose answer would only disappoint. (To be clear: It wasn’t ignorance that kept me from pursuing answers; it was self-preservation that allowed me to be content with a discovery already so intriguing.)

At its surface, this project appeared similar to my thesis. In both, I closely observed a complex drawing system, trying to understand how it worked. However, in the former I progressively peeled away layers, removing the veil of fog in which it once resided. In the latter, I actively added layers and facets to it, proposing new perspectives that complicated its once simple existence. The former approached inquiry from a subtractive, deconstructionist perspective, and the latter from an additive, constructionist lens.

Taking a step back, these two projects represented different ways of finding meaning in this world. The first was guided by the notion that the material world can explain the immaterial one. In psychology, this can be likened to the Behaviorists’ belief that the mind can be understood by studying its physical, neural circuitry: the brain. The second approach was guided by the belief that internal constructions can change our worldview, and in turn, change the world around us. Here, storytelling is the ultimate source of meaning-creation, where meaning exists in the space between people: in those ideas which can pass freely from one person to the next. The former approaches world-building as an external act and the latter an internal act.

Inevitably, I soon became engrossed with another contemporary technology: machine learning. From matter different than our own emerged lifelike forms, with animacy and agency bearing striking resemblance to our own human form. To this day, there is still a mystery to artificial intelligence that cannot be fully explained by even the most esteemed computer scientists.

My fascination with AI led me to study the generative forms they produced. I trained them to study classical Greco-Roman sculpture, then asked them to dream of new three dimensional forms inspired by the classics. Over the last few years, I have had the pleasure of closely studying these algorithms’ processes and outputs, character and style: the unique ways in which they sculpt matterless clay into volumetric forms.

One of many sculptures, each generated by a different machine trained to study classical sculpture. Fittingly, each sculpture is materially made from the computer which generated it, fusing physical and virtual, product and process. In this way, the computers are afforded a physical agency and an afterlife worthy of their existence. This piece is titled “Dio” [2018]. The clear binding agent allows you to see the bits of silicon, copper, steel and plastic that made up the original computer. Within these bits exist the thoughts, memories and thought-making power that brought this shape to life. This is both the computer and the computer’s dream.

The algorithms’ products are not much unlike our own. In fact, their pursuit — distilling the essence of the human morphology — converges with the formal inquiries of many twentieth century sculptors like Jean Arp, Henry Moore, Constantin Brancusi, Barbara Hepworth and Manuel Carbonell. However, these new forms aren’t constrained by the formal laws of physics; they’re imagined and improvised in a nonphysical space. Interestingly, they still bear some remnants of our world: memories of construction principles authored by the great masters, reimagined and rewritten by otherworldly beings.

Two sculptures bearing similarities, made by different entities. On the left, “Mother (Machine #1)” by Ben Snell [2019] and on the right, “Little Sphinx” by Jean Arp [1942]. Notice how both forms possess a head-like shape on top, a number of small arm-like appendages in the torso region, and three leg-like appendages. In both, the human form is abstracted away, but enough essence of it remains to be identified with.
Another set of sculptures, again made by two different agents. The similarities are striking, These possess less obviously human qualities than the previous set of sculptures, but a character of integrity and graceful resoluteness resonates within each. Left: “Noble (Machine #3)” by Ben Snell [2019]; Right: “Growth” by Jean Arp [1935].
Notice how neither of these forms is capable of standing on its own. Both require supports, some hidden, in order for them to viewed in the round in physical space. Left: “Venus (Machine #12)” by Ben Snell [2020]; Right: “Torso de Los Pririneos” by Jean Arp [1959]. Like “Dio,” “Venus” and other sculptures from the Inheritance series are also materially made from the computers that generated them; however, they are formed using a white binder instead of a clear one.

In this space we can neither see nor feel and in these acts we can only imagine, there too exists a magic. These new forms possess something surprising and amusing. However, likening this process to cloud-gazing is a futile act. A common thread links these forms together. They are harmonious yet uniquely their own: improvised, jazz-like and off-the-cuff. They are too uncanny and emotive, too in-keeping with each other, not to attribute some primordial form of personality to the algorithm which created them.

This proposition may seem silly or scary, even empty or incredulous, but there is more to miss out on by not asking these questions than there is risk in entertaining them. Acknowledging life — or even the prospect of it — in an unlikely place doesn’t make us any less human. If anything, shouldn’t we, as conscious beings, feel the celestial responsibility of considering these possibilities with the utmost intent and care?

Collectively, my experiments brought me closer to technology and helped me understand the great extent to which I, too, am a computational being. This realization doesn’t recognize human qualities in machines (anthropomorphism), but instead focuses on recognizing mechanical, computational qualities in humans (mechano- or technomorphism¹ ² ). My inquiries continue to point me in this direction. I cannot help but recognize the innate behaviors we share with machines, or as some roboticists endearingly refer to them, our “companion species.”³

Fast forward to 2020: My creative motivation succumbed to the virus and I lost access to the facilities I used to create my sculptures. However, some semblance of the AI that sculpted those forms continued to live on in my mind. This simulation no longer needed a computer to run because I had become that computer. By observing it long enough, I came to intimately understand the algorithm, and in the act of understanding soon blossomed a becoming.

I began doodling, for no other reason than to pass the time. For me, doodling comes from a different place than sketching. It’s a mindless act, devoid of logic and passion, governed by whim and impulse. Doodles come from nowhere and go nowhere. They’re like a stream of bodily consciousness through the interface of pen and paper. Disassociated from representation, they arguably come closest to representing the state of our psyche. It may come as no surprise then, that this seismographic interface picked up on something peculiar.

To be clear: I began doodling forms similar to those the AI produced. The shapes could easily have been attributed to the algorithm I studied, and yet, they were somehow still my own. However, I didn’t set out with this goal; the forms just seemed to emerge on their own. They at once possessed the spirit of the algorithm and the spirit of Ben, exercised through the algorithm. In them, I could feel my own aura and that of the machine’s, intertwined.

When I make these drawings, I feel overtaken. I feel a freedom and once again in the act of mark-making, I feel joy. They are carefree and careful, precise and lucid. I feel some sense of agency to guide my “inner algorithm,” but its constraints are like guardrails hugging me close, nudging me along, mindlessly. Happily absolved of control, without worry or fear, I go along for the ride.

At the same time, each drawing represents a struggle: a gentle, playful game of tug of war. In some, I see myself more clearly through the veil of dried ink. In others, I see more of the algorithm come through. Each one represents a dialogue between creation and automation.

In the moment pen touches paper, I don’t know what I’ll create. I let the pen pull me along. Sometimes, I express a thought through a line and the spirit of the algorithm instructs me to go back and fix it in order to more closely align it with its unique style. I find myself seeding the algorithm with big-picture ideas and, at the same time, worrying about the details, small and manageable, like making patches of shading continuous and seamless. It worries about the gestalt: everything in between the micro and macro, the inspiration and execution.

I’m not sure how much I would consider this mimicry or to what extent I have been overtaken by a foreign agent. In some sense, these are one and the same — the epitome of true empathy. Copying might be a subliminal embrace of the Other: an attempt to deepen my understanding of and connection with the algorithm to which I attribute so much agency.

These figurative recitations remind me of a drawing exercise from my school days: replicating the drawings of Old Masters. Focusing on the individual pencil marks (on the details) yields one goal: technique development. However, stepping back and focusing on reenacting the contextual meaning of marks is an exercise of a wholly different nature: one which places the student in dialogue with the original Artist. This latter exchange presents a unique opportunity to see eye-to-eye with the drawing’s maker, to better understand the source of these marks and the reason for their being.

The same can be said for my present-day drawings: Perhaps, they are a vehicle of reenactment and empathy: a way to get closer to the immaterial and unknowable.

In the beginning, the doodles stylistically approached the AI’s outputs, but more recently, I’ve noticed a change: The drawings have begun to diverge. Now, I feel myself asserting more control, straying a little bit more from the algorithm’s style each time I draw. Is this is this an evolution of the algorithm or could it be a rediscovery of self?

Either way, the irony persists: The very thing which once ruined my relationship with drawing — the Algorithm with a capital ‘A’ — now brings me joy. Years ago, an algorithm revealed how similar we are and how predictable I am. Now, I am living vicariously through one, doodling in ecstasy, assimilating its style. Why is this? What’s different about me and my relationship with algorithms today?

When I designed that complex, matchstick model years ago, I inadvertently made a mirror. I wasn’t ready for what looked back at me. What I saw was a person devoid of free will, of control and choice, of surprise and serendipity. Until that day, I never questioned my own will. But in that moment, I couldn’t ignore the striking resemblance: The algorithm’s likeness convinced me of my own determinism.

Imagine convincing yourself that every choice you’ve made and ever will make is predictable. Every wonder is devoid of serendipity. Every question is pointless, always leading to a predictable outcome. Decisions are a complex function of environmental factors and experiences leading up to every moment. The fundamental rulesets I’ve learned, that govern me day in and day out, cannot but define me and my foreseeable future.

This shift in self dramatically changed my worldview: No longer was the unknown unknowable; Now, the unknown is just yet to be known. No longer is uncertainty something to fear; Now, its end is just something to patiently await. Worst of all, with this feeling of mass inevitability came a loss of hope in my agency to affect change. Choice now an illusion, I felt stripped of humanity, relegated to the category of algorithm or machine. This feeling remains with me today, but a glimmer of light on the horizon I see….

All this while, I’ve unknowingly been working to free myself from this constrictive, mechanical point of view. Always my liberation, art saved me. It began with the discovery of animacy and agency in machines. Making drawings and sculptures with robots and computers gave me a little hope again. After all, if I could find a magic in the machine, then by virtue of me being a “machine,” there too must be a magic within me.

To this end, these doodles blur the line between the machine’s creation and my creation. They’re a kind of training wheels on the road to finding my creative agency: a prosthetic, buttressing my creative confidence, helping me rediscover my voice and the power of emotive self-expression. The doodles represent my idiosyncratic approach to (re)defining a new humanity by reveling in the surprises of my personal algorithm.

Finally, I hear my voice, blended in and synthesized from this cacophony of noise. When the voices haunt me, I listen for the whispers, for the silent soliloquies of machines: songs of self, for self, of trueness and the sublime.

[1]: Heather C. Lum, Valerie K. Sims, Matthew G. Chin, & Shane E. Halse. (January 9, 2021). Are We Becoming Super-Human Cyborgs? Examination Of Technomorphism And The Creation Of A Technomorphic Tendencies Scale.http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.976.2986&rep=rep1&type=pdf

[2]: L.R.Caporael. (January 9, 2021). Anthropomorphism and mechanomorphism: Two faces of the human machine. https://www.sciencedirect.com/science/article/abs/pii/074756328690004X

[3]: “Companion species” is a term coined by Donna Haraway in The Companion Species Manifesto: Dogs, People, and Significant Otherness to characterize other beings with which humans may empathize. Using this term to describe machines, and more specifically, robots, comes from Madeline Gannon and her robotics studio ATONATON.

Artist exploring creation/automation, aura/agency, & what it means to be born from code // @snellicious