The Accumulation Period

TAM-EAR.1-01 · The Standard with Ears · The Approximate Mind

A king named Hāla ruled a stretch of the Deccan sometime around the first century, though the century is a guess and the scholarship has pushed it back and forth across several hundred years. The dynastic lists put him seventeenth in his line. One of them gives him five years on the throne. He fought no war that anyone bothered to remember, and what survives of him instead is an anthology and a nickname. The nickname was kavi-vatsala, which means something like the one who felt toward poets the way a parent feels toward children.

The anthology runs to roughly seven hundred verses, each of them short, most of them four lines. They are not in Sanskrit. Sanskrit was the language his court conducted itself in, the language of its inscriptions and its learning and its gods. The verses are in Maharashtri Prakrit, and what they are about is villages. A woman watching a road. A buffalo in a river. Somebody’s mother-in-law. The particular misery of a house where two people are not speaking. Rivers get named. Trees get named. Rain arrives on ground that has been hot for months and the smell comes up off it, and that smell shows up again and again across the collection, in verse after verse, the way a thing shows up when everyone already knows it.

Almost none of the poets are named. Forty-four of the verses are attributed to the king himself, which is either true or the kind of thing that gets said about kings.

A court looking downward is ordinary. Courts have always done it, occasionally, when they wanted something. What is not ordinary is that we can count what the looking produced, and the number is seven hundred.

Seven hundred verses. That is close to the entire surviving record of what a living language was saying about itself, across generations, in a region of some millions of people. Everything else that was said in Maharashtri Prakrit by everyone who spoke it, all the arguments and jokes and instructions and complaints, every song that did not happen to be a four-line lyric about longing, every technique somebody worked out for doing something with their hands, is gone. Not suppressed. Not lost in a fire. Simply never collected, because collecting was hard and nobody had a reason.

The anthology is not evidence that the dark was penetrated. It is a measure of how little penetrating anyone could afford.

What the dark was doing

#

Culture differentiates through latency. Deviation accumulates where the standard is not looking, compounds there, and only then becomes something the standard has to reckon with. Why should the not-looking matter? If a swerve is good, why does it need the dark? Why not swerve in public?

Because a swerve in public is graded, and it is graded by the standard, in the standard’s terms, before it has terms of its own.

Consider what Prakrit actually was by the time Hāla’s anthologist got to it. Not broken Sanskrit. That is the reflex reading and it is the wrong one, and the reason it is wrong tells you what the accumulation period does. The Prakrits had their own phonology, consistent and rule-governed, describable, describable enough that grammarians eventually described them. They had their own prosody. They had genres, decorum, a sense of which subjects belonged in them and which did not. By the time these verses were composed there was already an idea of what a good gāthā was, held by people who had never written a line of Sanskrit and did not want to.

That is what the dark produces. Three things, roughly, and none of them are available to a deviation that is observed young.

The first is mass. One person saying a word a new way is an error. A thousand people saying it that way for sixty years is a language. Nothing distinguishes the two except repetition and time, and repetition and time are what the unobserved space supplies.

The second is internal rule, and it gets missed more often than the first. A deviation that has been running long enough starts developing its own sense of correct and incorrect, its own way of being done badly. Somebody in a village could compose a bad gāthā, and other people would know it was bad, and they would know it by a standard that had nothing to do with Sanskrit.

A deviation with rules of its own has stopped being a mistake and started being a rival.

That conversion happens in the dark or it does not happen.

The third is the ability to survive being dismissed. When the deviation finally meets the standard, and it always eventually does, it gets judged and the judgment usually goes against it. What determines whether it continues is not the judgment but whether there are enough people already doing it, with enough investment in doing it well, that the judgment simply fails to land. Prakrit was dismissed constantly. Sanskrit theorists had a whole vocabulary for its inferiority. It went on for centuries anyway, and eventually the languages that descend from it are what a few hundred million people speak.

So the function of the unobserved space is not concealment. It is deferred judgment. The dark withholds evaluation until the thing being evaluated has grown its own criteria, and a swerve judged before that point does not lose the argument. It never becomes the kind of thing that could have an argument.

Nobody granted it

#

There is a romance about the underground, and the romance is good. It says that the margins are where the real material is, that obscurity is a kind of integrity, that the people outside the institution are outside it on purpose and their being outside is what keeps them honest. It has produced a great deal of true observation. Hidden transcripts are real. The things people say when the overseer is not in the room are different from the things they say when he is, and the difference is where a lot of history lives.

But the romance smuggles in a premise that will not survive inspection, which is that the dark was in any sense chosen.

It was not. Nobody in a Deccan village in the second century decided to keep their language away from the court in order to let it mature. They spoke the way the people around them spoke because those were the people around them. The distance between what they said and what the inscriptions said was maintained by nobody. Bad roads produced it, and rare literacy, and expensive scribes, and the sheer physical difficulty of one person going to a place and writing down what they heard and bringing it back.

Every standard in history would have listened further down than it did if listening had been cheaper. Sanskrit did not decline to hear the villages out of restraint. Rome did not leave the provincial Latin of the fifth century alone out of a policy of linguistic pluralism; it left it alone because the administrative apparatus that had been holding the written form steady came apart, and once it came apart there was nobody to do the correcting. When institutions wanted to hear downward they built instruments for it, and the instruments were folklorists, censors, informers, tax assessors, priests taking confession. What all of those have in common is that they are people, that each one covers one place at a time, and that each one costs money.

Hāla, or whoever assembled the anthology in his name, is one of those instruments. He is the standard reaching down as far as its budget went, and the reach produced seven hundred verses, once.

The accumulation period was never a mercy. It was an expense no standard could meet.

The anthology complicates itself further. Scholars who work on the Sattasaī have argued that Maharashtri Prakrit as it appears there may already be a literary register, shaped for a court audience, closer to a stylization of village speech than to village speech. Which would mean that even the one time we get to look, we are not looking at the vernacular. We are looking at the vernacular dressed for the room it was carried into. The instrument does not just cost money. It changes what it collects, in the act of collecting.

The record of the ones that got out

#

Every case we have used so far shares a property, and the property is a problem.

Prakrit surfaced. Vulgar Latin surfaced, spectacularly, into half a dozen national languages. Those are the examples available because they are the examples that left evidence, and they left evidence by winning. The accumulation periods that ended in nothing left nothing, by definition. A vernacular that thickened for two centuries in some valley and was then extinguished by a plague, a conquest, a road that got built, or simply by the young people leaving, produced no grammar, no anthology, no descendant tongue, and no entry in anybody’s history of languages.

Which means the mechanism we have described is inferred from a sample selected on the outcome we are using it to explain. If long unobserved accumulation is what lets a deviation survive contact with the standard, the evidence for that is a set of deviations that survived contact with the standard. The ones that accumulated just as long and died anyway are not in the file.

We wonder how much of what we believe about the way culture changes is a description of the way culture surfaces, and whether the two have ever come apart from where any of us is standing.

The failure direction is asymmetric, which is what keeps this from collapsing the argument. We have no examples of a deviation that arrived fully formed, met the standard in its first season, was judged against the standard’s criteria, and went on to become a rival system anyway. Not one. The absence of that case is not proof, but it is the kind of absence that a claim can lean on. What it sets is a ceiling. Latency looks necessary. It is plainly not sufficient. Most of what accumulated in the dark, across all of human history, stayed there and then stopped.

So the nostalgia has a floor under it somewhere. The unobserved space was less a nursery than a place where the great majority of what grew did not make it out, and the few things that did make it out are the entire basis of our affection for it.

What is left over

#

So the condition has a shape. Deviation needs a period in which the standard is not reading it, that period has to be long, and nobody gives it. The period is a gap produced by the cost of attention, and it closes exactly as far as the cost of attention falls.

Which raises the question everything after this has to work through.

The obvious version is that the cost went to zero and the gap closed. Hearing is now continuous. There is no valley the road does not reach, no room without a microphone in someone’s pocket, no register that is not being deposited somewhere the moment it is produced. On that account the accumulation period is simply over, and everything that follows is an inventory of consequences.

That version is too clean.

Some things are still accumulating unobserved. Not many, and not the ones anyone would have predicted, and not in valleys. There are practices going on right now inside fully wired institutions, in rooms with network connections and logged systems and every device that a surveillance account would require, where somebody has been developing a way of doing something for twenty years and not one part of it has entered any record that anything reads. The hearing reaches those rooms. It reaches all the way in. Something else is going on, and it is not distance and it is not secrecy.

Whatever that is, it is not the old dark coming back. The old dark was a limit on the standard’s reach. This is something the standard reaches and passes over. Nobody is protecting those rooms. Nobody has claimed them, defended them, or noticed that they are open.

The dark that remains was not preserved. It was skipped.

What the mechanism is, and why the skipping falls where it falls, is a separate argument, taken up in The Standard with Ears.

What stays from the older story is how contingent the arrangement always was. Seven hundred verses is what a king who loved poets managed to save, in a life, out of everything that was being said around him. He is remembered for it. It was an extraordinary act of attention by the standards of his apparatus, and it captured a rounding error.

The rest went the way the smell of rain on hot ground goes, which is that it is overwhelming and then it is finished and then the ground is wet and ordinary, and the only reason we know it mattered to anyone is that somebody with a court and a scribe and an affection for poets happened to write down that it did.

This essay opens The Standard with Ears by establishing the mechanism the rest of the series tests: that cultural differentiation has always depended on a period of unobserved accumulation, and that the period was an artifact of cost rather than a protection anyone extended. The residue it ends on, the dark that is reached and passed over rather than missed, is taken up in The Standard with Ears, which names what does the passing over. It stands in conversation with The Counterparty in the main series, whose account of the uneven dark supplies that residue, and with The Closed Window, which arrives at a related one from the far end of the same mechanism.