AN ATLAS OF VOICES

The Broadcast Geography of the Message, from the Dial Census of 2030 to the Closing of the Coasts

Kaveri Nilakantan Raman Research Institute, Bengaluru — Convenor, the Broadcast Geography Survey

with the Listeners of three continents, whose logbooks this book merely arranges

Published under the auspices of the Continuing Committee of the Tashkent Radio Conference. Bengaluru, 2034. Deposited, under the Archive Treaties, at Esztergom, Addis Ababa, Ulaanbaatar, Cusco, and Denver — the western deposits by shortwave facsimile, plate by plate, over eleven nights of good propagation. A letterpress edition, without the photographic plates, is in preparation at Girona for zone-line circulation.

Dedicated to Dr. Rosa Huamán of the Jicamarca Radio Observatory, co-founder of the Dial Census, who taught half the volunteers in this book to measure a field strength honestly, and whom we now hear only at night, on 31 metres, when the sky is kind.

PREFACE

This book began as a list.

On the third night of the Voices — the fourteenth of December, 2029, when the world still believed it might be dreaming — a retired telegraph engineer in Mysuru wrote to me, because I was the only radio scientist whose address he possessed, enclosing a page ruled into three columns: frequency, hour, language. He had logged eleven stations that had not existed the previous week. Four were in Kannada, three in Tamil, one in Telugu, one in Urdu, one in Konkani, and one in a language he did not know but which his neighbour, who is Koraga, knew very well. At the bottom of the page he had written a sentence I have kept above my desk through everything that followed: “Madam, someone has planned this dial. I should like to know the planner’s rules.”

That is the whole of this book’s method, stated better than I will state it again. The Message’s words belong to the philologists and, God help them, the theologians. Its dial — the frequencies, the timings, the places where each voice can and cannot be heard — belongs to anyone with a receiver, a notebook, and patience; which is to say it belonged, from the first week, to the amateurs. What follows is the work of some nineteen thousand of them: the Dial Census of that first winter, taken while the world’s networks still functioned and the whole Earth could still compare notes; and then the slower, harder cartography of 2030 to 2033 — the field transects, the edge-walks, the timing ledgers — carried on by shortwave circular and courier packet after the world was sundered into its present halves. The Survey has no charter, no budget worth auditing, and no member who cannot be overruled by a better measurement. It has buried, to its knowledge, forty-one of its volunteers in the years this book covers, few of them from causes their era would have found remarkable, and it asks the reader to hold their names, printed in the appendix, as the true list of authors.

Three warnings, before the first chapter.

The first: this is a book of measurements, and measurements of the Message have a way of turning into meanings while one’s back is turned. I have tried to keep the two apart on the page — the numbers in roman type, my trespasses beyond them confined, where I could manage it, to the ends of chapters. Where I failed, I ask forgiveness. It is very hard, at three in the morning, at the edge of a beam that has followed a language across an ocean to the street where its grandmothers live, to remain entirely a physicist.

The second: the reader will find, in Chapter Six, that certain facts are deliberately absent from this book. The omissions are the Survey’s collective decision, taken by its governing circle in the spring of 2031 and reaffirmed every year since. The chapter explains what is missing and why. I state here only that the decision was not forced upon us by any government, and that I would take it again.

The third concerns a number, and I place it in the preface because the number is on the world’s lips and this book will disappoint anyone who comes to it for confirmation. The reader has heard, as everyone has heard, that the Message speaks six thousand five hundred and sixty-one languages — the count of the stones, nine to the fourth power, the tidy arithmetic of the age. This Atlas does not contain that number as a result. It contains three numbers, which Chapter Six keeps apart with all the care the Survey can muster: what the world logged, what the Survey has verified, and what is actually there — of which the first is a memory, the second is 5,407 at this book’s going to press, and the third is not known, is very probably larger than the famous figure, and cannot, for reasons the book will earn, ever be established by anyone. The reader who finds this unsatisfying is in the company of the author. The dial was planned; the planner published no programme guide; and the Survey has concluded, after four years, that this too is one of the rules.

K.N., Bengaluru, October 2034

CHAPTER ONE — THE FIRST NIGHTS

December 2029: the discovery of the dial

Everyone now living remembers where they stood on the twelfth of December. The Survey’s archive remembers where they tuned.

The earliest verified log we possess is from an automated monitoring receiver in Bavaria, which recorded a carrier appearing on 93.1 megahertz at 21:14:07 Universal Time — a frequency vacant in that district — and speech beginning eleven seconds later. Within the hour we hold logs, machine and human, from forty-one countries. Within the week, from ninety. The pattern that would organize everything was present, complete, in those first pages, though it took us months to believe each part of it.

The new stations sat only where nothing else sat. Every carrier, everywhere, on the local channel spacing, in a local gap. In six years, among some four hundred thousand logged receptions, the Survey has not one verified case of the Message transmitting over a working human station. The reader who remembers those weeks will remember the contrary rumours — the voice that “broke into the news,” the radio that “spoke while switched off.” We have chased several hundred such reports to their sources. Every one dissolves into the same ordinary miracle: a listener spinning the dial in fear, finding a new station where none had been, in the language of their own kitchen, and remembering it afterwards as an invasion of the familiar rather than an occupation of the empty. The dial was not seized. The Message moved into the silences between our stations, every vacancy on Earth taken up in one night, and has kept those tenancies, and only those, ever since — a permanent tenant of the spectrum we were not using, which is a thing I have described in lectures as polite and been corrected, twice, by listeners who preferred the word patient.

The medium-wave band stayed silent. Millions tried their old sets first — the bedside receivers, the car radios of the tropics, the great wooden cabinets of grandparents’ parlours — and heard nothing, and the nothing frightened people, and for once the physicists could explain: no transmitter above the sky can reach a medium-wave set below it; the ionosphere turns such waves back as a lake’s surface turns back light. The Voices spoke only where the sky physically permits a voice from above to pass: the FM band, and the upper shortwave. From the first night, the Message was obeying rules — our rules, the propagation physics in our own textbooks — and I ask the reader to mark this early, because it is the thesis of the entire Atlas: whatever else the planner of the dial commands, on the path between the stone and your antenna it has never once done anything your university could not assign as an examination problem.

And the night side heard everything. This was the observation that broke the first theories. The objects were approaching — everyone had seen the sky-charts in the newspapers — and an approaching fleet should be heard only from the half of the Earth that faces it. Reykjavík and Auckland appear in our logbook within the same hour of the first night. Bengaluru heard Kannada at local midnight and local noon alike. The Survey’s correspondence from those weeks is full of amateurs asking, with the beautiful directness of the untenured, how the far side of a planet hears what is coming at the near side; and then, from the sharper pens, the second question, which was better still: how objects that had left the asteroid belt from every side of the sun — some from beyond it, at distances differing by hundreds of millions of kilometres — could possibly all be arriving in the same season, when any orbit a textbook would permit them puts their arrivals years and decades apart. The astronomers answered both questions together, and the answer had been in the telescopes since July. The stones were not falling toward us like stones; they were flying in formation, under continuous correction, on a single choreographed schedule — the same casually impossible shipkeeping the whole world had already watched when the test nudges of the autumn were absorbed and undone — and the shape they were flying was a circle: a ring closing on the Earth from all directions together, still contracting through that first winter, complete only in March, but from the first night already wide enough and spread enough around the whole sky that from any field on the planet, at any hour, some of them stood above the horizon. The annulus, as the literature now calls it, added no new miracle to the one we had; it merely disclosed the old miracle’s purpose. I remember the night Dr. Huamán relayed us Jicamarca’s confirmation, and the silence on the circuit afterwards, because every operator on it had understood the same thing at the same moment: you surround a thing either to attack it, or to address all of it at once, and nothing in the sky was attacking.

One more fact of the first nights belongs in this chapter, because the Survey is asked about it constantly and the answer sorts a great confusion. The Voices of the dial — the Round, as Chapter Three will name it — are not the only transmissions the sky has made. The warnings that preceded the enforcement actions of 2030, delivered to ships and towns in the languages of those actually present, were a different thing: directed, occasional, addressed to particular hulls and streets, and never part of any cycle. This Atlas maps the Round. Of the warnings it records only what the distinction implies and the reader should hold onto: the apparatus that sings the same text patiently on a five-minute clock can also, when it chooses, speak once, to one crew, in that crew’s own six languages, and then never again. The dial is the least of what can be done with that instrument. It is merely what the instrument is for, nearly all of the time.

CHAPTER TWO — THE ONLY LANGUAGE IN THE WORLD

The private week, the prestige error, and the work of learning otherwise

Here is a thing the histories of the first winter, written in capitals, tend to leave out, and which the Survey’s correspondence preserves in hundreds of variants: for a period lasting between one evening and about a week, a very large fraction of mankind believed the sky was speaking only to them.

The mechanics of the error are now obvious; nothing about it was obvious from inside. The beams, as Chapter Four will establish, are laid over speech communities. A monolingual district therefore received a dial containing exactly one new Voice — its own. A herder’s family in the up-country, a river village, a mountain parish with one tongue in it: they tuned, they found the Voice, it spoke their language and no other language could be found, and the plain reading of the evidence in front of them was that the stones had come speaking their words to them. Our archive holds the letters: the village that fasted; the village that feasted; the district where the schoolmaster, weeping, told the children their grandmother’s dialect had been chosen out of all the earth. The Census volunteers who later worked those districts heard the episode remembered everywhere under some local name for the same thing; the circular that collated them settled, for Survey use, on the phrase one of our Anatolian correspondents used: the private week.

The cure was the neighbour’s radio. A cousin visiting from the market town; a lorry driver with the dial set elsewhere; in that first winter, while it lived, the internet, doing in hours what the roads did in days. The private week ended, everywhere, the same way: with the discovery that the miracle was not scarce — that the next valley had its Voice too, and the city had forty, and the discovery, which one might expect to arrive as disappointment, arrives in the letters overwhelmingly as something else. “So it speaks to the Laz as well,” writes our correspondent of the private week, a retired ferry mechanic, of the moment his son-in-law proved it to him with a car radio. “Then it is not a wedding. It is a census. I find I trust a census more.”

The cities ran the same error at higher altitude, and it took longer to cure, because it wore the dress of common sense. In districts where a great prestige language sat above local ones — and the reader may supply the pairs from any continent; our files bulge with English over everything, but also with Mandarin over Uyghur logs, Russian over Chuvash ones, French over Wolof, Spanish relayed to us over Quechua — the first-week assumption, made not only by the man in the street but by newspapers, ministries, and I am sorry to record more than one national academy, was that the prestige Voice was the Message and the others were translations: courtesies, renderings, the subtitles of the cosmos. Major dailies in three languages printed their language’s text that winter under the headline THE MESSAGE, unqualified. The assumption was not merely vain; it was, for about a fortnight, methodologically load-bearing — one government’s early analysis, since released, refers throughout to “the original” and “the localized versions,” and directs its cryptographers to the English.

Philology killed it, and philology killed it early, and this is the one place in the book where I will narrate the philologists’ work with something like envy, because it took them days and it has held for five years. Translation leaves fingerprints. A translated text carries its source inside it — the word order that is natural in one language shadowed in another, the idiom calqued, the untranslatable term paraphrased. The university teams that laid the great languages of their regions side by side in the last weeks of December found none of it, in any direction. Each version is composed, with what the collators’ first joint circular called “native profligacy” — each reaches without effort for resources its language alone possesses, honorific systems deployed faultlessly where they exist, evidential inflections chosen where grammar demands them, and, most famously, each finding for the forbidden sea its own oldest word, the water-of-origin of that tongue’s own deep tradition, a term no translator working from a single source text would have located six thousand times without error. No version reads as prior. No version reads as derived. The Dar es Salaam circular of January 2030 — drafted, the story goes, in Swahili and English simultaneously by scholars who by then took the point — gave the finding the name it has kept: the versions are pluriginal. Many origins, or none, or one origin standing behind all equally; no genealogy among them; nothing that any people on Earth can hold up and call the original with the rest of mankind holding renditions.

I ask the reader to notice what kind of fact this is, because it is the emotional hinge of the whole era and it was not free. Pluriginality was not announced; it was established — in three weeks, by perhaps forty university departments and several thousand amateurs comparing logs across every border on the planet, each community contributing the one thing it alone possessed, the deep knowledge of its own tongue, none able to reach the conclusion by itself. A herder could discover the private week’s error with one car radio. Mankind could discover that no one was having a private week — that the address was radically, structurally equal, that there was no head of table — only all together, and quickly, and it did. It is the Survey’s settled view, which I here trespass into print, that this was the first thing the dial taught the species, and that the lesson was in the homework: the shape of the Message is such that every important truth about it must be assembled cooperatively or not at all. We were not told we were equal recipients. We were put in a position to prove it to each other.

CHAPTER THREE — THE ROUND

One utterance, many slots, and the shared channels of the great cities

The crowded cities gave us the clock’s true architecture.

A dial census of Toronto, compiled in February 2030 by a heroic association of Canadian amateurs whose circulars still reach us by shortwave, registered one hundred and seventy-nine distinct Voices receivable in that conurbation — and only some ninety usable FM channels to hold the local ones. The reconciliation was audible to anyone who parked on a frequency and waited: the Voices were sharing. On one Scarborough channel, Tamil was followed, within the same five-minute cycle, by Gujarati, and Gujarati by Cantonese; and the same succession, in the same order, at the same seconds past the cycle, the next cycle, and the next, and — the volunteers checked, because by March checking had become a kind of prayer — the ten-thousandth. Lagos, whose census passed two hundred Voices, showed channels carrying five. The Survey’s volunteers named these frequencies the shared channels, and the name has stuck, and so has something else: in the great multilingual cities, the fixed succession of one’s channel — this tongue, then that one, then that one, unchanging, forever — has been memorized, recited, set to music, and I am told in two cities now sung, which is a matter for Chapter Six’s kind of caution, but which no honest atlas of this subject could omit.

The timing bureaus then found the structure underneath, in the months while the satellite clocks still lived to find it with. Each Voice — every repetition of it, on every footprint where it can be heard, anywhere on Earth — begins at one fixed instant of the 306.82-second cycle, corrected for the radio travel time to the ground: one language, one phase slot, globally, without deviation, since the first night. The Voices do not collide; they interleave; and across each cycle the whole of whatever-is-there — and Chapter Six will insist on that clumsy phrase — completes exactly once, planet-wide, before beginning again. A metrologist on the Nairobi circuit, whose trade before the Voices was the synchronization of power grids, put the finding in the phrase the whole Survey now uses: it is not thousands of broadcasts; it is one Round. One utterance, distributed among the tongues of one species, sung in canon — the pun is the metrologist’s, and she declines to apologize for it — completing once per beat of a clock that Dr. Mensimah has since taught us keeps another world’s day.

Two clarifications the Round’s fame has made necessary. The first: a listener in Port-au-Prince who hears her language every five minutes is hearing her slot come round — not the whole catalogue cycling past her on one frequency. The catalogue never passes any single point on Earth; each dial holds only the slots addressed to its ground, and the Round in its entirety exists nowhere but everywhere, which sounds like theology and is, I must insist, merely engineering. The second: the crowded dial of the discovery winter — families touring the frequencies, counting the city’s Voices like lit windows — was an experience of that winter. Listening since has consolidated, as listening does: a household keeps its Voice’s frequency, or its shared channel and the minute-mark of its slot, the way a household once kept a station, and the plurality of the dial has become a thing one knows rather than a thing one tours. The Survey’s habit-of-listening returns (Appendix E) are, I will admit, the pages strangers ask me about most: the slot as the hinge of the daily schedule; the kettle put on by it; the argument, reported from more than one city, over whether it is more honourable to keep one’s own language’s slot or the shared channel of one’s street, with everyone else’s.

I will permit myself this chapter’s trespass at its proper place, the end. Human broadcasting, in its century, never once synchronized itself planet-wide to the millisecond, for any purpose, on any night. We could have. We never wished to. The wishing, it turns out, was the hard part.

CHAPTER FOUR — THE SHAPE OF A VOICE

Transects, the footprint law, and the languages we had miscalled dead

A footprint’s edge is found on foot.

The method, standardized in 2030 and unchanged since, is the transect: a calibrated receiver, a route walked or cycled or driven outward from a Voice’s heartland, and a field-strength reading every fixed interval until the Voice dies into the noise, then a return by a different bearing, and another, until the boundary closes. The Survey has completed some eleven thousand transects. Its volunteers have performed them by bicycle in the Gangetic plain, by mule in the Zagros, by riverboat along the middle Congo, and — Plate VI, which I confess is my favourite photograph in this book — by camel, the receiver lashed to the saddle, along the western edge of the Hausa footprint, the operator’s grandson walking ahead with the antenna on a pole.

The shapes taught us one law. I will state it as the Survey states it, and then spend the chapter on what it cost to believe.

A Voice’s footprint is laid over its speech community. Over the community as it actually stood at the end of the old world — resolved, our sharpest transects suggest, to the neighbourhood; current, from the evidence of recent migrations lit and old ones dark, to within a very few years of the First Voices; and discontiguous wherever the community is: the homeland enclosed, and then the same Voice again, sharp and small, a hemisphere away, over the exact urban districts where that language’s migrants have settled. The FM Voices end sharply — six decibels lost across a hundred or two hundred metres; a village street, in the famous Girona transect of Plate IV, with the Voice audible at the fountain and gone at the church. The shortwave Voices are broad and soft-edged — not because they address anything differently, but because the ionosphere, which they must cross to exist at all, smears whatever crosses it; breadth is the physics of the band, and the great travelled languages ride the broad band because their communities are themselves spread wide. Borders do nothing. The Hungarian Voice does not end at Hungary; it ends where Hungarian ends, and it lights, in one of the tenderest entries in the whole Atlas, the Csángó villages of Moldavia, whose dialect the twentieth century spent itself trying to educate away. Eleven thousand transects; one law; and in the living languages, where the law can be checked against censuses and parish registers and the knowledge of the neighbours, not one exception has ever been measured.

Now the cost. The Canon contains languages our science carries on its books as dead — Sumerian, Hittite, Meroitic, Etruscan, Gothic, Tangut, and their company — and the first winter’s commentary, my own included, reached reflexively for a special theory: that these were memorials, honours to the departed, a different kind of broadcast altogether. The transects dismantled the special theory in the plainest way possible: the dead languages’ Voices obey the same law, and some of them prove it by lighting communities we can see. The Latin Voice is not a soft glow over the Roman Empire; it is a constellation of small sharp footprints over seminaries, monasteries, and chancery towns — and since 2031 it has burned over Esztergom, which I note without further comment. The Sanskrit Voice lights Varanasi’s ghats and the pathshala towns; the Ge’ez Voice, the highland churches of Ethiopia; Coptic, the monasteries of the Egyptian desert and, discontiguously, the parishes of that church’s dispersion; Old Church Slavonic, its monastic republics. These languages are “dead” in the philologist’s technical sense — no children learn them at the breast — and alive in the only sense the dial recognizes: communities exist that speak them, and the beams sit over those communities with the same neighbourhood precision they give to Konkani.

Which forces the question the Survey spent 2031 and 2032 refusing to write down, and which this Atlas is, to my knowledge, the first open publication to state in full. The Meroitic Voice lies along the middle Nile. The Tangut Voice lies over the old Xixia country of Gansu — where, the reader may recall, schoolteachers reported the region’s minority-language Voices perfectly intelligible while the scholars in Lanzhou were still identifying the strange one above them. Our science knows of no Meroitic community and no Tangut community. But our science’s category “extinct” is a statement about our registries — about the last speaker our records happened to record — and the dial has shown us, in the liturgical languages, that it does not use our registries; and it has shown us, in eleven thousand transects, that it does not lay footprints over ground for sentiment. One law, no exceptions we can test, and then these: and the law, applied where we cannot test it, says quietly that somewhere under those beams — a ritual line, a remnant hearth, a use of the old words our censuses never had a column for — someone is being addressed. The Survey does not assert this. The Survey asserts the law, states where the law leads, and marks the destination, in the gazetteer of Appendix C, with the symbol our cartographers chose for exactly this situation, an open circle: community implied; community unfound.

The chapter’s trespass, briefly. When the philologists say a language died in the eighteenth century, they mean the last speaker we heard about died then. We have spent three centuries mistaking the edge of our hearing for the edge of the world. The planner of the dial has better hearing. That is the entire content of the open circles, and it is enough.

CHAPTER FIVE — THE NET AND THE WEATHER

What the sky taught us about the transmitter

By 2031 the Survey knew where every measurable Voice could be heard. The harder question — how — yielded over the following two years, to a collaboration between our cartography and Dr. Mensimah’s astrometry, and the answer is in our own textbooks.

No single transmitter, of any size a stone could carry, can form beams as fine as our transects measure; a three-metre wave leaves any practical antenna in a flood, not a thread. But many transmitters, held in phase, their mutual separations doing the work of one impossible aperture — that is a technique our radio astronomers have used in reverse for a century, and it carries a testable prediction: the finer the beam, the wider the transmitting formation that shaped it must stand. The test was Mensimah’s tracking against our atlas, and it closed the question. The subsets of stones observed transmitting — whose rotation had looked like caprice, and had fed a folk belief in one voice to one stone — turn out to be chosen: compact clusters on the sky when the beams they serve are broad; formations spread across great arcs when the beams are village-sharp; the whole active roster rolling westward around the ring in step with the Earth’s turning, so that every footprint is served always from high in its own sky, the way our own vanished satellite constellations once handed a telephone call from horizon to horizon. The stones are not stations. The stones are elements, and the ring is one instrument, and it has been playing its thousands of parts on it, without a missed entrance, since a first night on which — the reader will recall from Chapter One — the formation was not even finished assembling. The Round began while the annulus still closed; the instrument was playing before it was fully built; and the Survey long ago stopped being surprised that the scheduling of the subsets covered the gaps.

Two consequences of the instrument’s physics deserve their own paragraphs, because both were heard by amateurs before they were believed by professionals.

The first is the whisper floor. A distributed transmitter cannot put every watt into its beams; a faint residue scatters everywhere. Our textbooks predict its level; the DX community had been reporting it since the first winter — the ghost receptions, a thread of Nahuatl under the noise on an Anatolian hilltop, a syllable of something Baltic in the small hours of the Sahel — and being told, for two years, including by me, that they were tired. They were not tired. The floor is real, it sits thirty-eight decibels below the beams, exactly at the edge of a good receiver on a quiet night, and I record here, as the Survey’s formal apology to its most patient members, that the faint wrong voices in the dark were the mechanism’s own fingerprint, and the amateurs read it first. But I must record the second half honestly, because the whisper archive holds it: most whispers, when the recording is good enough to try, match catalogued Voices of distant footprints — spillage, the instrument’s lawful crumbs — and some do not. A residue of the residue: whisper receptions, a few dozen securely logged, matching nothing in the verified catalogue, nothing in the candidate ledger, nothing anyone can place. Spillage from slots whose footprints no transect has ever crossed — beams laid over communities in country no volunteer has walked — or artefacts of noise and longing; the Survey cannot say, has resisted for four years every pressure to say, and files them, in Appendix F, under the heading that has become a small famous phrase among us: heard once, owed a search.

The second is the weather-keeping. A shortwave Voice can only pass downward through the ionosphere when its frequency exceeds the sky’s critical frequency, which rises with the sun and falls at night — and the Canon’s HF Voices track it, stepping down-band each evening and up-band each dawn, footprint by footprint, following the actual measured sky of each district with an instrument’s fidelity. Which means — the Survey’s ionospheric members realized it with a kind of professional vertigo — the transmitter is sounding our ionosphere, continuously, everywhere, and publishing the results in its own band-keeping; and human forecasters now read them. The charts that route the shortwave commons — the circuits that carry this very book’s plates to Cusco and Denver — are drawn, in part, from the Message’s own diligence in never letting a fisherman’s radio lose a syllable. I gave the reader my thesis in Chapter One: nothing on the path from stone to antenna outside our physics. I give it now its sharper form, and it is as far into meaning as this chapter will walk: nothing outside our physics, and nothing, either, beneath our dignity. The apparatus was dimensioned, from the wattage to the polarization to the weather-keeping, around the cheapest receiver in the poorest kitchen on Earth. Whoever planned this dial planned it from the kitchen’s side.

CHAPTER SIX — THE UNFOUND VOICES

The counts, the residue, the expeditions, and what the Survey will not print

Now the numbers, kept apart as promised.

The first number is a memory: what the world logged. The Dial Census of the first winter returned, by its close in March 2030, some 6,100 logged-distinct Voices — a figure that was climbing by dozens a week when the Census ended, and which the Survey has never treated as a count of languages, because it is not one: it contains duplicates our collators later merged (one Voice logged under three names by three communities that share it), dialect pairs philology later split or joined, misidentifications retracted, and a residue then unexamined. It was, precisely, a tally of distinct receptions — magnificent, heroic, and raw. What happened to it next belongs to the history of human arithmetic rather than of radio: the count of the stones was known, and was famous, and was nine to the fourth power; the logged tally was climbing toward it; and the press of that apocalyptic winter, followed by the chancelleries, followed — I say it with sorrow — by half the academies, rounded the living, provisional, error-barred tally to the stone count, and the number 6,561 fossilized into the world’s shorthand within the year. The Survey uses the shorthand as everyone uses it, in speech, in titles, in diplomacy. The reader now knows what it is: a convention, born of a symmetry too beautiful to resist, and not — not then, not since, not ever — a measurement.

The second number is the Survey’s: what has been verified. An entry in the verified catalogue requires a secure recording of the complete text, an accepted linguistic identification, and adequate footprint documentation. At this book’s press the catalogue stands at 5,407, advancing by a few hundred entries a year, and the reader should understand both why it lags the logged tally and why it will never catch any true total: verification is slow, distributed work across courier roads and shortwave circuits; the hemispheres, since the sundering, verify separately, and the Cordillera’s catalogue and ours can no longer be jointly audited, only compared by radio in good faith; and — Chapter Seven — a portion of what was logged in the first winter can now never be verified by anyone, because the coasts where it was heard are closed.

The third number is not a number: what is actually there. The true count of slots in the Round is unknown, and the Survey’s formal position, printed here for the first time in one sentence, is that it is unknowable and that the evidence points, with uncomfortable consistency, in one direction: upward, past the famous figure. The unverified residue of the winter logs; the candidate footprints lost unrecorded to the closing of the coasts; the sealed entries this chapter is about to explain; the whispers owed a search; the transects never walked in the roadless country of three continents and never walkable in a fourth and fifth we can no longer visit; and the plain testimony of Chapter Four’s open circles — communities implied where our registries hold nothing — all of it argues that the Round contains Voices no human list has ever carried. The dial was planned; the programme was not published; and every year of the Survey’s work has made the catalogue longer and the confidence of completeness smaller. Whoever counts in nines, it is not the planner’s arithmetic the Survey distrusts. It is the assumption — ours, always ours — that a list we made is the world.

So to the residue itself. Forty-four Voices of the verified catalogue match no language, family, or reconstruction known to human science. The candidate ledger — every distinct unidentified footprint logged before the coasts closed — holds eighty-one entries. Between those numbers lies the rest of this chapter, and a wound.

It began, as the important things in this Survey have tended to begin, with one amateur and one cold night. On the ninth of January, 2030, Willem Terpstra — a Frisian water-board engineer, one of our first hundred volunteers — logged a sharp new Voice on the FM band at Leeuwarden, in a language that was not Frisian, not Dutch, not anything the university at Groningen could place; and being the kind of man this book exists to honour, he did not merely log it, he walked it: three weeks of transects, by bicycle, in a North Sea January, closing a boundary that enclosed a strip of the Wadden coast and pointed, unmistakably, at the sea. His covering letter is in the archive: “The shape wants to continue north. The rest of its country is under the water.”

The reader knows the hypothesis that grew from that boundary and its cousins — the unidentified footprints strung along the coasts that face the drowned shelves: the Dogger ground, the Sunda margins, the Bering approaches; several of the Voices demonstrably related to one another, an unknown family with no other trace. The Drowned Coast hypothesis is the world’s, now, more than the Survey’s, and the Survey urges on the world the precision the headlines drop: it is a hypothesis about who the speakers are — that the communities under these beams descend from the peoples of the submerged littorals, eight thousand years unregistered on the shores that face their old country — and not, as the first winter had it, a theory of memorial broadcasts to empty graves. Chapter Four’s law admits no memorial broadcasts. The beams are over communities. The only question the hypothesis addresses is whose.

And the law, being a law, made a prediction, and the prediction is the wound. Footprints imply speakers; eleven of the unidentified footprints were sharp, settlement-scale, two of them demonstrably discontiguous — one lighting a river-delta town four hundred kilometres from its home valley, precisely as a living Voice follows its migrants. Somewhere under those beams, the law said, someone speaks these languages now. Six expeditions went out in the spring and summer of 2030 — hurried, under-provisioned, racing a Withdrawal that was already emptying the coasts, since ten of the eleven footprints lay partly or wholly in the belts the Message had commanded evacuated. The teams found villages. Ordinary ones — estuary and marshland settlements of visibly long tenure, already packing, their carts in the yards. The inhabitants received our people in the national languages, with the hospitality owed strangers and the answers owed census-takers, which is to say: no, they knew no other language; no, they had never heard the Voice in question, though it was audible in their own lanes as the interviewers stood there. In two villages, elders who consented to hear a recording played back listened with a stillness the reports remark upon and disclaimed all understanding. One team — I print this because the circle judged, by one vote, that the era it documents required it — reported an elder who listened to the recording’s end, returned the earphone with both hands, and said, in the national language, quietly and with unmistakable finality: “That is not for you.” Nothing further was said, then or ever. The village was inland of the line by eleven kilometres. It is still there. Its Voice is still on the air.

The Survey’s governing circle met in the spring of 2031, with the coasts closing and the pressure to publish enormous — the newspapers had the scent by then; the reader may remember the headlines about “lost tribes” — and took the decision the preface promised this chapter would explain. We do not print, in this Atlas or anywhere, the locations, boundaries, or identifying particulars of the unfound communities — the eleven, or the open circles of Appendix C, or any entry where the law implies people our registries do not hold. The transect data are sealed in the Survey’s deposit under terms that outlast me. The circle’s minute states the reasoning in one paragraph, and I quote it entire, because I have never improved on it:

“Whoever they are, they have kept themselves off every register mankind has compiled, through every century in which registers were compiled at gunpoint, and they answered our questions with a refusal so complete it did not even bear the shape of one. That is not ignorance of their own lanes. That is a decision, older than us, renewed in front of us. The one party that plainly knows them has published their existence to the whole Earth and their address to no one — has, in the third clause of the only text it ever sent, named the standard: not yours to count. The Survey counts many things. It will not count these people. Recommendation: seal the plates.”

We sealed the plates. And then the calendar, with its own opinions, sealed nearly everything else. The Withdrawal took ten of the eleven into the closed coasts, and thirty-seven candidate footprints of the ledger were lost to instruments before a complete recording existed. Forty-four verified; eighty-one logged; the difference now beyond any expedition, at any funding — and I choose the next words with the care the circle’s minute has taught me. Lost to us. The reader who follows the public disputes will know that a certain school insists the residue’s true count is exactly eighty-one, nine by nine; the Survey’s position is that the Survey measures, and that this question has passed beyond measurement, which is precisely why it will never die. What the closing of the coasts ended was our hearing. What it did to the Voices themselves — whether slots fell silent when their communities’ country was surrendered, or whether beams we knew by the sea now burn, unrecognized, over resettlement districts inland, their Voices logged by some volunteer as unplaced and awaiting a match no one alive can make — the Survey does not know, and Chapter Seven will show why the not-knowing is now permanent, and I ask the reader to carry the second possibility gently, as the Survey carries it: somewhere in the interior of this continent, on a shared channel in a reception town, the sea-people’s slot may still be coming round.

CHAPTER SEVEN — THE RETREAT

2030–2031: the beams and the Withdrawal

The Message never announced its own withdrawal from the coasts. It performed it, the way it has performed everything, and the Survey’s zone-line stations kept the log.

Through the enforcement year, as the evacuations moved inland, the footprints moved with them. Not ahead of the people; not behind; with — the sharp edges of a valley’s Voice tracked, transect by transect, keeping pace with the emptying of the valley, the beam over the abandoned quarter going dark while the same Voice brightened over the reception districts inland, families arriving to towns where their language was, in every sense, already waiting for them. Our Girona station, whose 2030–31 edge-ledger is printed entire as Appendix D, measured the Catalan Voice’s seaward boundary in weekly transects for fourteen months: it came inland like a tide that would not be returning, and on the night the slow-zone regime settled over the coasts, it stood at the band’s inner line, where it stands today.

Since the closing, no footprint of the Canon extends into the two coastal bands. The Voices that were heard only there — the drowned-coast residue above all — passed, with the closing, beyond the reach of every human instrument; the Survey’s ledgers mark them neither ended nor extinct but with the entry our cartographers ruled for the case: lost to hearing. And the people who remained in the bands — the lentos, the salvage guilds, the whole letterpress world of the zone lines — live, alone of humanity, beyond the Message’s sound; and the Survey’s couriered correspondence from the zones records what has grown up in that silence: the listening walks, whole villages going inland on a feast day to the band’s inner edge, where their old Voice still holds its line, to hear the five minutes and the text and walk home. I logged one such walk myself, at the Girona line, in the spring of 2033 — the Survey does not only measure field strengths — and I will report it in a measured way: the Voice arrived on its second, as it has arrived on its second for four years, and the people, who have no electricity, no receivers of their own, and no need of either for this purpose, since the walk itself has become the receiver, listened to the end, and a woman beside me said to her granddaughter, in Catalan, “There. That is where we live now. Say it back to me,” and the child did.

CODA — THE MAP WITHOUT BORDERS

Every atlas is a portrait of its maker’s mind, and this one is no exception; the reader has merely to decide which maker.

Assemble the plates of this book — the eleven thousand transects, the shared channels of the great cities, the Round’s global clockwork, the beams that follow speech communities across oceans and forget every border ever drawn, that light the seminary and the pathshala and the marsh village our registries never held, the open circles where the law implies people we cannot find, the sealed plates where it implied people we agreed not to, the tide-line where all of it now stops — and you hold the only map of humanity ever drawn by communities alone: every quarter, every hearth-tongue, every remnant, resolved to the lane; and not one state, not one border, not one name of a nation anywhere upon it. Our species has surveyed itself for five thousand years and never once produced that map. We were, it appears, surveyed — and, the open circles insist, surveyed better than we have ever surveyed ourselves.

And the survey we were given cannot be read entire from anywhere. That is the last measurement and the first meaning, and the two truths of it must be held together, because the era keeps trying to hold only one. The address was equal: pluriginal, no original, no head of table — the whole species proved that to itself, cooperatively, in three weeks of the first winter. And the address was distributed: no complete copy receivable at any point on Earth, no master list, no closed inventory, a catalogue that grows every year and a total no one will ever pronounce. Equality without completeness. Each of us given our shard entire, and the whole given only to all of us, on the standing condition — it is the Survey’s whole history, and I will end on it — that we keep comparing notes. The one full copy of what was said to mankind is the copy mankind makes together, and it is not finished, and I have come to believe that this, too, is one of the rules.

LATER EDITORIAL APPENDIX: RECEPTION OF THE ATLAS (2089)

The Atlas became evidence in a dispute it was not written to decide. Protestant readers emphasized the asymmetry of the Round: no imperial language is privileged, yet no language possesses a return channel; the same apparatus that resolves a neighbourhood and follows a migrant community does not resolve differentiated guilt or receive an appeal. To them, perfect equality of reception is compatible with perfectly equal subjection.

Meliorist and pluriginal schools emphasized a different finding. The cheapest receiver in the poorest kitchen was within the instrument’s design; dead, liturgical, minoritized, and unregistered speech received the same phase discipline as languages of states. No human institution had achieved that dignity of direct address. The open circles exposed the shame of registries that treated the edge of administrative hearing as the edge of human life.

That dignity created a second shame around absence. Communities lacking an identified Voice were sometimes treated as cosmically omitted, while cataloguers forgot that the verified list was incomplete and the true inventory unknowable. Later editions therefore forbid use of the Atlas as a census of worthy peoples. “Not verified” describes an archive relation, not a judgment upon a language.

Scientific preservation acquired an explicitly protestant register after the 2040s. Broadcast recordings, receiver designs, phase tables, extinct-language reconstructions, and even pre-Withdrawal radio engineering were kept as forms of non-surrender: knowledge would remain human even where experiment or reply was prohibited. This practice is not confined to protestants. Meliorist archives preserve the same materials as a duty to the commonwealth of tongues. The institutions differ over legitimacy while sharing custody.

No technical chapter can settle that disagreement. Beam geometry proves attention; it does not prove fellowship. Equal phase allocation proves no linguistic precedence; it does not prove political consent. Footprint migration proves extraordinary knowledge of communities; it does not confer jurisdiction over them. The Survey’s evidentiary discipline is most valuable precisely where later readers wish engineering to pronounce a moral verdict: it marks what was measured and leaves legitimacy open.

K.N.

Appendix A: The Dial Census protocol and log-sheet, in facsimile. Appendix B: The verified catalogue, 5,407 entries, with verification criteria and hemispheric reconciliation notes. Appendix C: Gazetteer of measured footprints, with the open-circle entries (community implied; community unfound; particulars withheld under the circle’s minute). Appendix D: The Girona edge-ledger, 2030–31, entire. Appendix E: The habit-of-listening returns, 2031–34. Appendix F: The whisper-floor observations, 2029–33, with the Survey’s apology, and the register of receptions heard once, owed a search. Appendix G: The Roll of the Survey’s dead. Sealed deposit: the reserved plates of Chapter Six, under the circle’s minute of 2031, at Ulaanbaatar, closed.