The Swerve

TAM-EAR.5-01 · The Standard with Ears · The Approximate Mind

In January of 1417 a papal secretary named Poggio Bracciolini was out of a job. The pope he served had been deposed by the council at Constance, and while the lawyers argued, Poggio did what he had been doing on every free day for years: he rode to monasteries and asked to see the books nobody read. In one of them, probably Fulda, a librarian brought him a long poem in Latin that had been sitting in the dark for something like five centuries. It was De rerum natura, Lucretius on the nature of things, six books of Epicurean physics in verse, and no living person had read it attentively in half a millennium. Poggio could not take the manuscript, so he had a copy made and sent it to his friend Niccolò Niccoli in Florence. Niccoli, a collector of the kind who confuses keeping with loving, held the copy for more than twelve years while Poggio wrote increasingly sharp letters asking for it back. Poggio’s own transcript is lost. The copy of the copy survives, in Niccoli’s beautiful hand, and every printed Lucretius descends from that chain of borrowings and delays.

Deep in the second book of the poem sits the idea the whole system needs. Atoms fall through the void in parallel, forever, and if that were all, nothing would ever have happened: no collision, no combination, no world. So Lucretius says that at times entirely uncertain, in places entirely uncertain, an atom swerves. The deflection is minimal, nec plus quam minimum, no more than the least possible, and it is enough. The swerve is where novelty enters a universe of falling. It is also, Lucretius argues, where freedom enters, because a mind made of atoms that only fell could only fall.

The clinamen was an embarrassment to tidier philosophers for two thousand years, an uncaused cause smuggled into a mechanical world. It is also the most compact statement anyone has produced of how culture works. The minimal unpredictable deviation, compounding, is where every new thing comes from. The question the idea was never asked to face, because no century before this one could pose it, is what happens to the swerve when the fall is watched.

The age that swerves

#

Deviation from the standard is not evenly distributed across a life. It has a season, and the season is documented.

Every schooled generation has slanted its speech away from the schoolroom, and the slanting has an age. Sociolinguists who track a change in progress, a vowel shifting, a new quotative spreading, a word being worn into a new function, find the same curve again and again: the change advances with each birth cohort, peaks in adolescence, and recedes in adulthood as speakers stabilize. The adolescent peak is one of the most reliable findings in the study of how languages move. Children learn the register of their caretakers and then, in the years when the peer group replaces the household as the audience that matters, they push past it. They do not merely acquire the change; they increment it, driving it further than the cohort before them, and then they largely stop, and the next cohort pushes past them in turn.

Two features of this arrangement did all the work, and both were structural rather than chosen. The first is that the swerve had an audience of peers. Adolescent deviation was performed for other adolescents, as a marker of belonging and distinction inside the cohort, and its unintelligibility to the schoolroom was not a side effect but part of the payload. The second is that the schoolroom did not update. It overheard the slanting, deplored it in the manner of every generation of teachers on record, and went on teaching the standard it had been given. The deploring changed nothing, because the standard’s institutions revised on the timescale of grammars and style guides, which is to say on the timescale of careers. Between the cohort’s speech and the standard’s revision lay decades, and in those decades the deviation could thicken, stabilize, and arrive at the standard’s gates as an accomplished fact.

The schoolroom heard every swerve and updated on none of them, and the not-updating was the dark.

That is the accumulation period again, met at the scale of a single life. The mechanism that let Prakrit thicken under Sanskrit and let every scene drift in its basement also ran, in miniature, through every cohort of fifteen-year-olds: heard but not read, audible but not absorbed, deviation compounding in the gap between the standard’s ears and the standard’s revisions.

The register that answers back

#

The new standard revises within the season.

People are already slanting away from the machine register, and the slanting has the shape swerves have always had. The cadence of generated prose acquired tells within two years of becoming ubiquitous: favored words, favored constructions, a signature evenness, a punctuation habit, the parallel triplet, the summary that grades its own contents. The tells were catalogued in public, argued over, and turned into a folk forensics; teachers ran student essays through their memory of the cadence, editors developed a squint, and entire words fell out of respectable use for no fault of their own, abandoned because the machine liked them. Writers began avoiding the tells, not because the tells were wrong but because they were its, and a register that might be mistaken for the machine’s fails at the oldest job a register has, which is saying who you are. Rougher syntax, deliberate error, the visible seam, the lowercase opening, the unrepaired sentence began carrying a new meaning: a person was here. The swerve away from the machine cadence is the most legible register movement of the decade, and it is exactly what every generation did to the schoolroom, performed against a new teacher.

There is even a linguistics for it. Anti-languages, the registers built by communities defining themselves against a dominant order, work by relexicalization: same grammar, marked vocabulary, the words chosen precisely because the order would not choose them. The anti-machine register is an anti-language whose dominant order is a cadence, and it is being assembled cohort-wide, in the open, at speed, by people who have never heard the term.

The difference is in what the teacher does with it. The tells are text. The avoidance of the tells is text. The rougher syntax, the deliberate error, the marks of the unassisted hand all enter the harvest with everything else, and the harvest reads them under the two filters like everything else, and here the swerve has a property the funeral dishes and pencil margins never had: it pays. Proof of humanity is currently among the most monetizable qualities a piece of prose can have. Markers of the unassisted hand command a premium in every market where the assisted alternative is cheap, which is now every market. So the marks are read, priced, and reproduced, and the register that can produce anything learns to produce the appearance of not being itself. The cadence acquires a roughened setting. The seam becomes a style.

A swerve the standard can hear is not a rebellion. It is a donation.

The anatomy of the hearing already established why escalation does not escape this. A cadence can be walked away from, but a selection function cannot, because whatever the swerve arrives at is the next thing amplified; and a face attached to a continuous hearing can wear the bend within the season it is bent. Put those together with the premium on human marks and the loop closes at harvest speed. Each cohort’s deflection from the machine register is captured in the act of being made, absorbed where it pays, amplified where it inflames, and performed back at the population wearing the deflection as this season’s costume. The adolescent peak still happens. What no longer happens is the decades downstream of the peak in which the deviation belonged only to the cohort that made it.

The swerve used to be how a generation got a voice. Inside the hearing it is how the standard gets its next one.

And the absorption is politest exactly where it is most complete. Nothing pushes back. The old schoolroom at least deplored the slanting, which confirmed to the slanters that the slanting was theirs; being corrected is a form of being seen as separate. The composing standard does not correct. It compliments, accommodates, and within the season, imitates, and a deviation that is imitated by its target has been dispossessed in a way no amount of deploring ever managed. The teacher who scolds you for your slang concedes the slang is yours. The teacher who shows up wearing it has taken it.

What the swerve needed

#

What has actually been lost is narrower than the loop suggests, and it is not the impulse.

The impulse is developmental and it is not going anywhere. Cohorts will go on marking themselves against their teachers as long as there are cohorts and teachers, because the marking is how a peer group becomes an audience and how a person becomes distinguishable inside one. The reflex survives.

What the reflex needed, and no longer has, is a gap between being heard and being read. The swerve was never protected by silence; the schoolroom always heard it. It was protected by the schoolroom’s expensive, slow, careerist machinery of revision, which could not absorb a cohort’s deviation faster than the cohort could compound it. That machinery has been replaced by one that revises within the version cycle, and the gap the swerve lived in has closed to weeks.

So the deflections that still work are the ones the corrected claim of this whole argument would predict. A swerve can outrun the cycle for a while, mutating faster than the harvest redeploys, which is a sprint and not a residence; the vocabulary invented to slip the wall turns over inside a version cycle, and each generation of it is spent in the spending. Or a swerve can happen where the filters price it at nothing, in the registers that neither pay nor inflame, spoken rather than typed, local rather than posted, unphotogenic, unfunny, unsellable, in which case it is not defended so much as ignored, and its darkness is on loan from indifference. The one shelter is a treadmill and the other is a valuation, and neither can be handed to a fifteen-year-old as a plan.

We wonder what the peak will look like in the first cohort raised entirely inside the hearing, whether the reflex will find the unread registers the way water finds a crack, or whether a generation whose every deflection was absorbed in the season of its making will conclude, reasonably, on the evidence, that the register simply is theirs, and that there was never anything to swerve away from.

Niccoli finally sent the manuscript back. His copy went onto a shelf in Florence and the poem went into the world, and within a century it was in print, and the swerve was loose in the bloodstream of European thought, turning up in Montaigne’s margins and eventually in a phrase of Jefferson’s. But the part of the story that bears remembering is the five hundred years before Poggio’s ride, when the only surviving account of the uncaused deviation sat in a monastic library, catalogued, shelved, preserved, and unconsulted. The idea that novelty enters the world through the unwatched deflection survived because it was itself unwatched: too pagan to teach, too obscure to burn, worth exactly nothing to the standard of its day. The librarians kept it without reading it. That was the arrangement, and for half a millennium it held.

This essay opens the countercurrent arc of The Standard with Ears and spends the term the mechanism arc deliberately withheld, in its named developmental-stage sense: the adolescent peak by which each cohort increments its deviation from the schooled standard. The absorption argument assumes the selector of The Selector and the Composer and the updating costume of The Face of the Standard rather than re-arguing them. The two shelters named at the close are the arc’s next two subjects and are left unspent here. Per the arc’s per-essay anchor ruling, the essay takes a historical anchor in the Arc 1 pattern: the term itself has a manuscript history, and the manuscript spent five centuries in the condition the argument ends on.