Infinite Scroll, endless repetition
I’m browsing through an album of physical photographs from my childhood. In one photo, two little girls smile in Easter dresses and floppy hats, posed in front of a cherry blossom tree. In another, the girls sit in front of the television, lapping up bowls of Campbell chicken noodle soup in fuzzy pajamas. In a third, the girls are sitting in front of a computer screen playing a 1999 computer game called American Girls Premiere, scripting scenes using avatars from the American Girls pantheon of characters.
Without thinking, I pinch my fingers outward on the photograph to get a closer look at the computer screen in the photo. I pause and squint, confused. I’ve been so habituated to interacting with photos on my iPhone that I’ve tricked my brain into examining physical artifacts in the same way.
Over the twelve years I’ve owned a smartphone, my fingers have been performing a familiar choreography: swipe, scroll, pinch, tap, double tap, drag and drop. My right thumb does the heavy lifting – tapping out text messages, scrolling through Twitter, swiping away notifications – but it’s only one part of the the suite of interactions that make up the gestures and touch events developers code into their apps. In one popular library for detecting gestures, there are 39 different types of interaction events developers can employ.
Learning this dance requires constant repetition. Tap, double tap, swipe. Photos are constantly being regenerated as I make new memories, automatically populating my phone’s photo library. My iCloud photos only began in 2017, when I started paying for cloud storage. All my other photos are scattered across disparate digital storage spaces – folders in Dropbox, external hard drives stuffed in a drawer, Facebook albums and tagged photos that disappeared when I deleted my account. (Those photos probably still exist somewhere on some Meta server, helping train an image recognition model).
In her essay "The Enduring Ephemeral, or the Future Is a Memory," Wendy Hui Kyong Chun argues that our digital tools have effectively “combined the transitory with the permanent, the passing with the stable,” thus making “the permanent into an enduring ephemeral.” Our memories become less stable as our digital photo album’s size increases.
The promise of digital media was that our memories would last forever, prolonging the life of physical photos that tear and fade over time. The reality is that the overwhelming abundance of digital media makes such memories more ephemeral, not less. All of my photos from college are effectively inaccessible, hidden in some digital grotto somewhere.
This conflation of memory and storage is no accident, says Chun. She says that the “everyday usage and parlance arrests memory and its degenerative possibilities in order to support dreams of superhuman digital programmability.” In other words, the degenerative quality of our digital media means that we are constantly forgetting and re-membering ourselves. Ultimately, the promise of digital technologies is transformation: The ability to program and reprogram oneself over and over again.
By the time the global pandemic hit in 2020, I’d grown especially tired of sitting in front of a screen. Like lots of other tech workers, I spend all day hopping between different windows, from Zoom calls to browser windows, from my terminal to a text editor. Desperate to disconnect, I began to shift some of my creative practice offline – off the grid.
I began to work with a vintage Japanese knitting machine from the 1980s. Quickly, my fingers learned a different kind of choreography to operate the knitting machine. I taught myself how to loop thin strands of yarn across hundreds of tiny silver needles, one needle at a time, in order to cast on or cast off a knitted piece.
I started first with the Brother KH-840, which relies on punch cards for knitting patterns in an analog way. I eventually moved on to the Brother KH-930e, an electronic knitting machine which relies on a circuit board instead of punch cards for storing patterns. Each subsequent version of the electronic knitting machine arrived with a slightly greater memory capacity, for storing a larger and larger library of knitting patterns.
This type of knitting machine was constructed specifically for the purpose of efficiently storing and communicating data. Like early computers, these vintage Brother knitting machines operate using punch cards. Punch cards, or pattern cards, had been used for weaving in Jacquard looms since 1801. Punch card technology eventually influenced Herman Hollerith’s development of the electromechanical tabulating machine for processing large amounts of data, an ancestor of the modern-day computer.
Punch cards in the knitting machine work as follows: Data is stored on each card as punched holes. In this binary system of 0s and 1s, a punched hole corresponds to one color of yarn and a non-punched hole corresponds to a second color. To knit the pattern, the card is run through the knitting machine, line by line, which has a set of mechanics that can physically detect the location of the holes. That data is then communicated, line by line, to the needle bed.
To my surprise, the process of knitting on a bed of needles had resonances with my experience as a programmer. Both knitting and coding are repetitive and procedural: Creating a pattern on a punch card requires manually punching out holes according to a finite set of specifications. Once the pattern is set, I run the knitting carriage over a bed of needles, repeating the same back-and-forth motion hundreds of times. It’s like writing a ‘for loop’ in code: a set of instructions or tasks that the script repeats over and over until it’s finished cycling through an array of elements. When something’s gone awry and a stitch drops off one of the needles, I pick up the stitch from the bottom with a special tool that hand knits the stitch vertically upward. This is the type of ‘debugging’ required to fix errors on the knitting machine.
Working with my hands in this way – from punching out holes in the punch card, to casting on stitches, to running the carriage – gives me a tactility that I found was missing from my digital practice. At the same time, in the process of disconnecting from one grid I found myself zeroing into another kind of “grid”: The online knitting machine community.
There are only a handful of YouTubers who make instructional videos about how to use a knitting machine. Because knitting machines are no longer being manufactured, very few of us have access to a vintage machine unless we are lucky to find the rare eBay listing. As I learned how to use my machine, I became intimately familiar with the voices of these YouTubers, most of whom were women. I didn’t always see their faces, but began to recognize their fingers as they moved in repetitive motion.
First there was @theanswerladyknits, whose step-by-step tutorials I watched over and over again. She taught me three different ways to cast on and off stitches and to shape a neckline. Then there was @girlyknits, whose instructions on knitting a bra top helped me learn to knit a folded hem and mock ribbing. @KnitFactoryImpl shared complicated mathematical formulas for calculating stitches and rows for a knit, teaching me how to draft my own patterns. Finally, I learned how to hack my Brother machine with the help of a YouTube video posted 12 years ago.
Some of these YouTube videos had fewer than 50 views, which made me curious about the women who appeared to be fiercely dedicated to documenting: What were their lives like? Why did they make these tutorials? For whom? Did they expect anyone would watch them?
The history of knitting communities goes back centuries. Critically, industrial knitting machines completely disrupted knitting guilds, which for centuries had built community and cultivated consensus around the quality of knitting. The very first mechanical knitting machine was a stocking frame, invented in 1589 by William Lee near Nottingham in Britain. The Luddites resisted the transition to factories by destroying these stocking frame machines. Luddites were protesting not just mass industrialization, but also the poor, mass-produced quality of textiles.
I suspect this small but highly active network of knitters, hackers, and artists on YouTube represents an attempt to recapture that sense of community, a guild of knitting machine enthusiasts. The community of women I stumbled across on YouTube has somehow endured over time: I find myself glued to the screen as I inch through video tutorials that had been posted years ago, painstakingly studying tiny finger movements.
In her essay on memory, Wendy Hui Kyong Chun says that "...we must analyze, as we try to grasp a present that is always degenerating, the ways in which ephemerality is made to endure. What is surprising is not that digital media fades but rather that it stays at all and that we stay transfixed by our screens as its ephemerality endures."
The dream of digital media has always been programmability: The ability to make and remake oneself over and over again. But as I unravel the yarn from yet another failed attempt at a mohair sweater, working my fingers through hundreds of stitches that will need to be re-knit, I realize that my attempt to find an offline outlet has simply plugged me into another kind of matrix. This version also holds the promise of re-generation – remaking oneself through infinite, repetitive movements.