For years I held a conviction I could not shake: that memory and language do most of our thinking for us. The words available decide what can be said, the memories to hand decide what comes to mind first, and together they lay down the routes a thought will take before it has properly started. If that was so, then an original thought, one that did more than rearrange what was already there, would need some distance from both. It seemed worth testing on the only mind I had access to.
So I set out to unlearn recalling, and I trained my mind the way one trains a dog: telling it, over and over, what to do and what not to do, rewarding it when it complied and scolding it when nothing had changed. Looking back, I was trying to escape habit by drilling a new one. Slowly, and then more than I had bargained for, things began to slip. Ordinary words arrived late, or not at all. Tasks I had done for years needed a few seconds of reconstruction before my hands knew what to do, and whole stretches of what I knew went missing.
Something else came with the losses. With no autopilot to carry me through the mundane, I had to attend to everything, even the routines I had stopped noticing years before. I was rediscovering what I already knew. It was slow and inefficient, and there was wonder in it. There was also embarrassment, the scramble to work out on the spot something I was supposed to know already, in front of people who assumed I did, and sometimes I failed. I reversed the experiment because I could no longer move among other people with the ease I had depended on. Undoing it took far longer than doing it, and some of what went has never come back.
Autopilot, I came to see, is memory doing its work out of sight. Much of what we perceive is supplied by what we already know, and recall fills in the scene before we have properly looked. A century ago Frederic Bartlett asked students in Cambridge to read a Native American story, "The War of the Ghosts", and to retell it from memory at intervals. The retellings changed in many ways, with omissions, substitutions and inventions, and the changes leaned one way: canoes became boats, the hunting of seals became fishing, and the ghosts, which made no sense to the students, faded or were explained away. Memory rebuilds as it recalls, and it compresses, keeping what fits the rememberer's world and letting much of the rest go. The wonder I felt was like the world arriving before it had been made familiar.
Compression is older than words. A cat remembers its territory, its routes, the neighbouring houses it likes to visit, and it marks that territory with scent other cats can read, a claim folded into a smell that says who passed and how long ago. A honeybee folds a flight into a dance: the angle of its waggle run gives the direction of the flowers and the run's duration their distance, and the bees that follow it can fly there. Whether any of this counts as language depends on where the line is drawn, and the line has kept moving.
Human languages compress in their own ways, through habits of description that shape what is kept. An Aymara model of time, documented in the Andes, places the past in front and the future behind, because what has happened can be seen and what has not cannot; to have a memory ahead of you is to have it in view. Speakers of Guugu Yimithirr, in northern Queensland, describe space by the compass, and in experiments they remembered the arrangement of objects by the compass too, where Dutch speakers remembered left and right. Habits of description and habits of attention can grow together, and over a lifetime they help shape a point of view. Which language becomes the official one, in a school or a colony, settles whose way of compressing the world passes for plain fact, and the struggle over who keeps a shared past runs through that settlement.
Oppressive language does more than represent violence; it is violence; does more than represent the limits of knowledge; it limits knowledge.Toni Morrison, Nobel Lecture, 1993
What lies between people is a bridge of a kind, and it cannot take the full weight of anyone's experience. Language is what lets us walk it without breaking it. It reduces what we have lived to a load the crossing can bear, and in the reducing it makes something the experience did not have before it was said, a shape another person can take up. Maurice Halbwachs argued that we remember within the frames others give us, and the shared shorthand is part of those frames. During the experiment, more of the world demanded my attention, and less of it came readily into words. The embarrassment came each time someone handed me something compressed and I had lost the means to open it. The constraint I had tried to break and the bond I could not live without turned out to be the same compression, seen from the two ends.
What I keep from those months is the wonder, along with a few gaps that never closed. I had wanted an original thought, and what I found was the familiar made strange, which is something else, though it may be where original thoughts start. The wonder, at least, can be had without forgetting. A familiar reading can be interrupted for a while, deliberately, and then allowed to settle again so it can be shared. Another language does some of this for me now. This site runs in three languages, none of them treated as the original, and every piece arrives in the others a little changed: a sentence in the Carrying method that ends on "taking" in English ends in Spanish on despojo, dispossession, a word that names what the English leaves the reader to infer. Moving between them, I find more difference than I ever found by subtracting. I still try to arrive at familiar things before autopilot gets there first.