Note on the translation — This translation re-fixes the source text without displacing its substance. A few terms of the corpus have no obvious equivalent in English; the choices made, and how they are to be read, are set out in the section “Translation choices” at the end of the volume. This translation is not frozen: it will evolve with readers’ feedback and with the improvement of the means of translation. The French version is authoritative.

Threshold

For a Processual Approach to the Movement of the Datum

Tome 2 — From Spring to Torrent

The datum can be given: circulation is its frequent and structuring regime, not its universal destination.

The word carries it: datum — from which French takes donnée — is what is handed over to someone, for a later occasion. An inscription can leave its birthplace: circulation is a frequent and structuring regime of it, not its universal destination; when a transmission is aimed at, it is carried by an intention, present or deferred, towards an identified recipient or towards a public to come. This tome has this movement as its subject: the putting into circulation of the datum, which extends its availability and its possibility of use and carries it to the edge of a reader.

From this movement, the tome brings out a guiding thesis: the datum navigates because it is fixed. Far from being opposed, its fixity and its journey condition each other — one transports only what holds. And from this thesis the rest follows: an act is needed to put a datum into circulation, as one was needed to draw it; transport is not the movement of a thing, but a relation between two movements, that of the datum towards the reader and that of the reader towards the datum; transporting a datum does not move it, but multiplies its carriers; and to propagate the datum is never to propagate knowledge, which requires a share that transport cannot supply in the reader’s place.

The eight documents that make up this tome lead this idea from the act of exchange to its runaway — from the spring, that stable point which humans plant in the flow, to the torrent, that regime in which accelerated circulation produces more data than any gaze can follow. At the end of the journey, the tome reaches a threshold: what arrives without being read remains mute, and what is decided there — whether a reader actualizes it or not — no longer belongs to circulation. It belongs to reading, and that is the subject of the next tome.

This tome therefore stops at the edge of the reader. It says what it is true to say of the movement of the datum for a long time; it leaves to others, and to the reader, the task of determining what to do with it for today.

The Datum Navigates Because It Is Fixed

Double Navigation, and the Paradox of a Fixed Point That Travels

Tome 2 — Document 1

“A datum is what is given — for a later occasion.”

— Alfred North Whitehead, Process and Reality (1929)

“Water that does not flow becomes impure.”

— after Ovid, Epistulae ex Ponto, I, 5, 6 (ut capiant uitium, ni moueantur, aquae)

Preamble — A Datum That Does Not Circulate

The first tome closed by announcing this one: Digging the Canal had opened up inter-entity transmission and handed its unfolding over to the tome of movement. Here it is. What the first tome established — the making of the datum, from drawing to fixation — holds here as an established result; only what movement requires will be recalled from it, step by step.

But a datum made and never transmitted remains solitary. The entry inscribed in the notebook of a single observer, never shown, never sent, never consulted, exists — it has been inscribed, it is fixed — and yet it serves nothing and no one other than the one who produced it. It is like water that does not flow: it exists, but it goes foul without having irrigated anything.

Now humanity does not make its data to keep them to itself. It transmits them. Circulation is a frequent and structuring regime of the datum, not its universal destination; when an inscription aims at a transmission, this gesture is carried by an intention — present or deferred, towards an identified recipient or towards an anonymous public to come. This is what the language preserves in the word itself: the French donnée comes from the Latin datum, “what is given”. The datum, etymologically, is not what is independently; it is what has been handed over to someone. Whitehead, whom the essay cites, said it precisely: a datum is what is given, for a later occasion. The datum calls for a recipient; without this horizon, the inscription remains a suspended gesture, waiting for what comes next.

A Tome That Goes from Spring to Torrent

This second tome has this movement as its subject: putting into circulation, which extends the availability of the datum and its possibility of use. Its title states its route: from spring to torrent. These two terms need, from the outset, to be situated in the network of images of the first tome — the river, which is the flow of the real; the canal, which is the flow of observation that humans dig in it, and whose locus is Digging the Canal. The spring, in this network, is not an object one would find ready-made: between the threshold of the instantaneous and the threshold of the eternal, humans plant their springs — the stable points from which they will draw data whose real flows slowly enough for the use they make of them (The Two Times of the Datum). The tome starts from there: from the datum at its spring, at the point where it is emitted. And it goes as far as the torrent — the regime that circulation reaches when it accelerates and aggregates to the point of carrying everything away. From spring to torrent: this is the route of the movement of the datum, from its point of emission to its runaway.

But movement cannot be described without first resolving a paradox — a twin of the paradox that The Two Times of the Datum has already met, and which threatens, as long as it is not dissolved, to make the movement of the datum unintelligible.

The paradox is this. The first tome insisted at length on the fact that the datum is frozen. Once inscribed and fixed, it no longer moves: it is even from this that it derives its usefulness, for one can transmit, consult, replay only what remains identical to itself. And now this tome announces that it navigates, that it travels, that it moves from place to place. How can what is frozen navigate? How can a fixed point travel?

The thesis of this document fits in one sentence: the datum navigates because it is fixed. Far from being opposed to movement, fixity is its condition. It is not despite its fixity that the datum travels; it is thanks to it. Establishing this requires distinguishing two navigations that usage confuses, as the first tome distinguished the two times of the datum.

Part I — Why Fixity Makes Movement Possible

What Does Not Hold Cannot Be Transported

Let us return to the established definition: to produce a datum is to arrest a value and fix it on a medium — the two operations that The Two Times of the Datum established. From that instant, the datum is withdrawn from the flow: it abides.

It is this permanence that opens up the deferred operations that The Two Times of the Datum listed: consulting, replaying, duplicating, transmitting. None of these operations is possible on what does not abide. One does not transmit a value that has already dissolved. One does not transport a vibration that has died out before arriving. Transport presupposes that there is, all along the route, something to transport — that is, a value that holds.

Fixity is therefore not the opposite of movement: it is its material condition. Just as one can ship a parcel only if it keeps its shape during the trip, one can make a datum travel only because it remains identical to itself from one end to the other. What comes apart on the way does not arrive; what does not arrive has not travelled.

The Fixed Point and Its Medium

The Two Times of the Datum had offered an image: the datum is a fixed point planted in a current — its strength and its limit hold together in this position. The present document adds a dimension that the first tome did not have to consider.

This fixed point can be made to travel. But let us be precise here about what travels. The datum is not in its medium: it is carried by it — The Two Times of the Datum established this in connection with duplication. What moves materially is therefore never “the datum” as a thing one would lift; it is the medium that carries it: the paper letter, the scroll, the disk, or else the signal that a wave modulates. To make a datum travel is to make a medium that carries it travel — or to produce a copy of it elsewhere, that is, to have it carried by a second medium without ceasing to have it carried by the first.

And it is precisely because the datum is fixed — because it does not dissolve when the medium that carries it is handled — that it can be moved in this way without being lost. A current does not move: it is the movement. A fixed point lets itself be carried, because it keeps its shape while it is carried. Here is the first navigation: the movement, from one place to another, of the medium that carries the datum. We shall call it spatial navigation. But there is a second one, more discreet, which the first tome has already described under another name.

Part II — The Two Navigations

Spatial Navigation: Travelling on the Medium

The first navigation is the one that can be seen. The medium that carries the datum leaves its place of origin and goes to another place, or else a modulation propagates along a medium in place. The letter leaves the sender and reaches the recipient; the scroll crosses the Mediterranean; the signal runs along the fibre. In all these cases, the datum moves in space, carried by a medium that makes the trip.

This navigation is visible, measurable, datable. It has a point of departure and a point of arrival. It takes time, sometimes negligible, sometimes considerable. It follows a material path, whether this path is a road, a cable or a wave. It is navigation in the most immediate sense: a movement from one point in space to another.

The tome now opening is, essentially, the examination of this spatial navigation: what triggers it, what carries it, what constrains it, what accumulates when it is repeated. But it is not the only one. If one kept to it alone, one would miss something that the first tome has already brought to light, and which “travels” with the datum without any medium being responsible for it.

Temporal Navigation: The Real Moving Away around the Datum

The Two Times of the Datum established it, and the formulation is subtle: the datum does not move — what moves is the real, whose flow pursues its course and moves away from the point the act had marked; lapse is not a movement of the datum, but a relation between a motionless datum and a mobile real.

The Two Times of the Datum gave a decisive image of this reversal, which must be retained for this whole tome: where classical science measures the movement of an object in a fixed frame, the datum presents for observation the moving-away of a mobile frame around a fixed object. This tome adds one word to it, and this word is important: around. The real does not move away from the datum in a straight line, as one moves away from a point by walking straight ahead. It turns around it while moving away from it — like the moon which, at each revolution, remains bound to the earth, still acts on it, and yet moves imperceptibly away from it. The datum remains in relation with the fragment of the real it drew: it continues to say something about it, as the moon still acts on the tides. But this relation slackens, turn after turn, and the gap grows irreversibly. The movement of the datum in time is not a rectilinear flight; it is a widening spiral, a bond maintained and a growing distance.

This is temporal navigation: not the datum moving, but the real going away while turning around it. It requires no transport medium, no trip, no path. It takes place on the spot, through the mere flowing of the real. A datum locked in a chest, never transmitted, navigates in this way in time without moving a single step in space. The Two Times of the Datum described it under the name of lapse; the present document names it temporal navigation in order to set it alongside the other — for the two navigations, as we are about to see, are not independent.

What the Two Navigations Have in Common, and What Separates Them

The two navigations share the same condition: fixity. Spatial navigation is possible because the datum holds while the medium that carries it is moved. Temporal navigation is possible because the datum holds while the real moves away around it. Without fixity, neither the one nor the other: what dissolves arrives nowhere and bears witness to nothing.

But they part company on one point. Spatial navigation is intended: it proceeds from a human act that decides to move the medium — what the next document will call the act of exchange. Temporal navigation has its source in a flowing that no one commands: the flowing of the real is not decided. In this first sense, it is done to us rather than done by us.

Let us beware, however, of concluding that humans are entirely passive in it. For if the flowing itself cannot be commanded, humans nevertheless have holds on the gap. They act on the real — The Wave and the Ocean showed that humans are a portion of the flow, and this tome will show that the torrent is the real marked by us; impacting the real means acting on the speed at which the fragment drawn moves away from the datum. And they have another hold: they can, through a new drawing, produce a fresh datum alongside the old one — not to make up the latter’s gap, which can never be recovered, but to reconstitute, through cadenced accumulation, a flow that follows the real. Humans do not abolish temporal navigation; they modulate its effect, to a certain extent. The exact extent of this hold — how far humans can hold the gap, and to what end — goes beyond this tome, which confines itself to noting that it exists: temporal navigation has its source in an undergone flowing, but it is not, for all that, beyond all human action.

This asymmetry has a consequence that the whole tome will unfold. If spatial navigation takes time — and it always does, however tiny —, then while it takes place, temporal navigation continues. The datum that its medium makes travel in space ages in time during the journey. It arrives further from its real than it was at departure. Transport is therefore itself a cause of lapse: to move a datum is to let it age for the time of the movement. The two navigations add up, and it is their addition that will make, at the end of this tome, the whole problem of the torrent.

Part III — What Double Navigation Illuminates

Why Speed Becomes an Issue

If the datum made only the spatial navigation, its speed of transport would be a mere convenience: arriving sooner would be better than arriving later, without anything essential depending on it. But because temporal navigation runs during transport, speed ceases to be a convenience and becomes an issue of substance. The slower the transport, the more the datum has aged on arrival; the faster it is, the less it has moved away from its real. The speed of transport is, literally, a struggle against lapse.

This is what sheds light on the deep meaning of the acceleration of transmission techniques, which Digging the Canal documented. Transmitting faster is not only a gain in comfort: it makes the datum arrive less aged, closer to the real it draws. At the limit — that of near-instantaneous transmission — transport ageing almost disappears: the datum arrives as fresh as it left. Digging the Canal had identified a systemic leap there, and its nature can now be measured: it is not that transport has become faster; it is that its own duration has fallen below the threshold of perception, and with it the ageing it imposed. We shall return to this at length; it is enough here to note that it is double navigation that makes this issue thinkable.

Why a Transported Datum Is Not a Rejuvenated Datum

A confusion must be set aside from the start, for it is persistent. Transporting a datum quickly makes it arrive little aged; it does not rejuvenate it. A datum drawn ten years ago and transmitted today in a fraction of a second arrives intact, but it remains a ten-year-old datum. The speed of transport acts on transport ageing — the delay of the trip — not on ageing of origin — the gap between the instant of the drawing and the present.

This distinction is crucial, and double navigation makes it possible to hold it. Spatial speed corrects only the temporal share that it itself adds. It does not go back up the course of the time of the real — for The Two Times of the Datum established it: freshness can never be recovered; the act that could have captured it is past. That is why a torrent of data transmitted at high speed can be, despite its transport freshness, a torrent of data already stale at the source: speed has made them arrive fast; it has not made them current. The tome will return to this point when it examines what accelerated circulation does — and does not do.

Why the Datum Is a Means of Not Letting Everything Be Carried Away

One last remark opens the way to the rest of the tome, and gives the movement of the datum its deepest stake. The first tome opened on a sentence: we do not live surrounded by data, but by a real that flows. The real carries everything away; nothing that passes abides of itself. Now humans, a portion of the flow, have found, through a deliberate gesture, the means of not letting everything be carried away: they inscribe. Taking a photograph of this present is making a fragment of it abide when the flow has carried everything away; it is tearing a share from the current, however tiny, and fixing it.

The datum is this means. It is the instrument by which one portion of the real — humans — holds another portion against the flowing that, without it, would erase it. And if the datum must be fixed in order to travel, it is because this means is fully valuable only when shared: one holds the present not only for oneself, but in order to transmit it — to another place, to another time, to another than oneself. It is through this that humans have built themselves as humans, that they have been able to accumulate, transmit, evolve: not by preventing the flow from flowing, which is impossible, but by fixing and circulating holds on this flow. The fixed point is fixed in order to be able to be given; we meet again the etymological clue: the datum is what is given for a later occasion.

But the later occasion — the encounter with a reader, the act by which they will decipher what reaches them — does not belong to this tome. This tome leads the datum to the threshold of this encounter, and stops there. It shows how the fixed point is set in motion, how this movement is engaged, unfolds, accelerates, and ends up producing more data than any reader will be able to receive — to the point that the very means humans had given themselves for not letting everything be carried away threatens, turned around, to carry everything away again. What then becomes of what arrives without being read — a datum actualized by a reading, or a datum that has remained mute — is the question of the next tome. The present tome’s sole charge is to bring the datum within reach. That is already a great deal, and it is what the documents that follow will establish.

Landmarks

For the references used in this document — the philosophical precision of the word datum in Whitehead, the Bergsonian distinction between duration and spatialized time that grounds the two times of the datum —, see the Annotated Bibliography of Tome 2. The methodological status is set out in the Preface to the Edition; the letter, the scroll, the fibre, the chest, the photograph, the spiral of the moon belong to the status of illustrations. The concepts taken up from the first tome refer back to their loci: the two times, fixation and its deferred operations, duplication, lapse and the springs planted between two thresholds, to The Two Times of the Datum; the situated reader and the caused/intended divide, to The Trace and the Reader; the flow of observation, the systemic leap and the acceleration of transmission techniques, to Digging the Canal.

This document carries, for the trilogy, the locus of double navigation. The thesis — the datum navigates because it is fixed — is demonstrated here; spatial navigation and temporal navigation receive their names here, and with them the image of the spiral — the real moving away from the datum while turning around it, a bond maintained and a growing distance; the distinction between transport ageing and ageing of origin is established here, as is the addition of the two navigations — transport itself a cause of lapse — of which the torrent, at the end of the tome, will make the whole problem. Everywhere else, these established results circulate without being demonstrated again.

The Act of Exchange

The Second Gesture, and the Medium That Carries It

Tome 2 — Document 2

“To give, to receive, to return.”

— Marcel Mauss, Essai sur le don (1925)

Preamble — Symmetry of a Gesture

The Two Times of the Datum established that the datum is born of two operations: an act that brings it into being — the act of inscription —, and a medium that makes it last — fixation. The first anchors the datum in the time of the real; the second sets it in the time of inscription. The datum is the tension between these two anchorings.

But a datum inscribed and fixed exists only where it was produced. For it to exist elsewhere — with another person, in another institution, in a time other than that of its place of origin —, a second gesture must intervene. The present document calls this gesture the act of exchange, and it shows that it is strictly symmetrical to the act of inscription.

The symmetry between these two acts is exact, and it is the thesis of this document. The act of inscription engages a fixation medium; the act of exchange engages a transport medium. The act of inscription brings the real to the status of datum; the act of exchange brings the datum from one point to another. The act of inscription founds the datum; the act of exchange extends its availability and its possibility of use. This symmetry is not an ornament: it is the key to a large part of what is at stake in the social use of data.

A precision of vocabulary governs what follows: clearly distinguishing the act that decides the putting into circulation — the act of exchange — from the trip it triggers — transport, the subject of the next document. The act is one-off and intended; transport is extended and material. Not confusing them is the condition for thinking correctly about what comes next.

Part I — A Second Founding Gesture

Why an Act Is Needed, and Not a Mere Event

One might believe that the movement of the datum goes without saying: that once inscribed, it “circulates” by itself, as water flows down a natural slope. This is an illusion, the same one that made people believe that the datum is picked up in the real like a ready-made object. The Trace and the Reader dispelled it for inscription: between the legible real and the datum, nothing crosses the threshold on its own; an act is needed that decides, that aims, that arrests. The present document dispels it for movement: between the fixed datum and the transmitted datum, nothing is crossed on its own; an act is needed that decides to make it pass.

This act is of the same nature as the act of inscription. It is human: someone decides to send, to publish, to share, to make available. It is intentional: one transmits for something, towards someone or towards a public, with an aim. It is situated: it takes place at an instant, in a place, with the means of an era. And it is, like the act of inscription, dated and unrepeatable: the sending of this datum, at this instant, towards this recipient, is a unique event of the flow of the real; one can perform another; one cannot perform that one again.

The distinction that The Trace and the Reader established — the trace is caused, the datum is intended — finds its extension here. A datum that “leaks”, that spreads without anyone having decided it, is not the object of an act of exchange: it is an accident, a causal effect, the equivalent for movement of what the trace is for inscription. The act of exchange is intended. There is someone who performs it, and therefore someone who answers for it.

Cost and Responsibility

Like the act of inscription, the act of exchange has a cost. Transmitting is never free: it requires a medium, energy, time, an infrastructure. The next document will show that this cost never disappears, even when it becomes so low that one stops seeing it. For now, let us retain that an act that has a cost is an act one chooses to perform or not — and this choice commits the one who makes it.

For the act of exchange engages a responsibility, and one with a reach of its own. The act of inscription already engaged the responsibility of the drawing — what to observe, with what probity, leaving what out of frame: The Trace and the Reader established it. The act of exchange adds to it the responsibility of putting into circulation: to whom I transmit, what I make accessible, what I choose not to put into circulation. One can have inscribed with probity and transmit without discernment; one can have drawn an accurate datum and put it into hands that will misuse it. The responsibility of exchange is not covered by that of inscription: it is a distinct responsibility, which is born with the second gesture.

That is why the movement of the datum is not a neutral technical fact. Deciding to transmit is performing an act one answers for — on the same footing, and in a different way, as one answers for the drawing one has made.

Part II — The Transport Medium

Every Transport Requires a Material Medium

The act of inscription holds only through a fixation medium: a material that keeps the inscription over time. The act of exchange, likewise, is accomplished only through a medium — but a medium with another function. It is no longer the medium that makes last; it is the medium that carries across. The present document calls it the transport medium.

Every transport requires a material medium. This is true for goods, for people, for words; it is also true for data. A datum that travels always takes a vehicle. The papyrus scroll crosses the sea in a crate, carried by a ship and by hands. The clay tablet passes from one city to another through a messenger. The book is conveyed by post with other objects. In each of these cases, the transport medium is visible: one can point at it.

A Sometimes Invisible Materiality

The contemporary illusion consists in believing that the digital datum, for its part, is not transported materially. The essay here takes as established — without demonstrating it — that a digital datum that travels is, at every instant, carried by a physical medium: the fibre that modulates a light signal, the antenna that emits a wave, the submarine cable that carries photons, the buffer memory that briefly holds the packet before forwarding it. Under this condition, the illusion comes from the speed and invisibility of transport — not from its absence; and digital transport is as material as postal transport, only faster, more distributed, harder to see.

This precision is not a technical detail. It opens up the question of the material conditions of transport — energy cost, infrastructures, dependencies, vulnerabilities — which are as real for the digital as for older media. A digital datum has a transport footprint; it is never purely immaterial. To believe the contrary is to make, about movement, the error the first tome fought about existence: taking for immaterial what is only invisible.

A Minimal Typology

Transport media can be described in three families, which correspond to three ways in which materiality ensures movement. This typology is minimal and belongs to substance; the concrete forms that populate it are dated.

Media that are moved: the scroll, the book, the tablet, the letter, the memory stick. The medium is itself mobile; it carries the datum with it and delivers it on arrival. The datum and its medium make the journey together.

Media that transmit: the voice that carries speech, the wave that carries the signal, the cable that carries the current. The medium is in place, fixed or stationary; what travels is a modulation of this medium — vibration, wave, pulse. The datum passes through the medium without the medium moving.

Media that make accessible: the wall that bears the fresco, the monument that bears the inscription, the screen that displays, the shelf that presents the book. The medium is not designed for movement; it offers the datum to whoever comes to it. Transport, in this case, is not that of the datum towards the recipient, but that of the recipient towards the datum — a point that the next document will develop at length.

These three families are not watertight. A lent book is a moved medium; read on the spot in a library, it becomes a medium that makes accessible. A broadcast signal is a medium that transmits; captured by a receiver, it is inscribed on a fixation medium at the recipient’s end and returns to the regime of the fixed datum. The same case can therefore combine several media depending on the moment of the trip being considered.

Part III — What the Act of Exchange Illuminates

The Complete Grid: Three Acts, Two Media, Conditions

One can now state the complete grid of processual ontology, as the two tomes carry it together. At the origin, the flow of the real. On this flow, an act of inscription draws a value; a fixation medium keeps it. The datum exists. Then an act of exchange decides to make it pass; a transport medium conveys it. The datum circulates.

Three acts, then — but the third is held back, and this tome does not deal with it: the act of reading, by which a recipient will make an actualized meaning from the arrived datum. And on its side, the symmetry of media stops. Fixation has its medium; transport has its own; resumption has no one medium — it has conditions of interpretive resumption: a reachable occurrence, a living code, a practicable interface, a reference corpus, competencies, a community, a body. No homogeneous medium provides this whole; the grid therefore cannot be forced into three equal columns. Their loss has degrees, of which the tome of the reader gives the scale. Conceptual parsimony invites one not to multiply acts without necessity; the tome keeps to exchange and leaves reading to the next tome. The grid it completes stops at the threshold of the reader — faithful in this to the dividing line between the two tomes.

Why Maintaining a Flow Is So Demanding

This grid sheds light on a point that Digging the Canal had left implicit. A flow of observation — that apparatus in which inscription is repeated to reconstitute the movement of the real — is not only a repeated inscription. It is a chain of maintenance in three operations — the chain of an apparatus, distinct from the grid of founding acts —: the drawings are inscribed, then transported to a place of storage, then made available to whoever will consult them. Inscription, transport, making available are repeated in parallel.

That is why tending a flow of observation is so demanding: it is not one but three chains that must be kept in working order. One can inscribe perfectly and transport badly; transport faithfully and make access impracticable; ensure everything upstream and let storage degrade. Each link has its own failures, and the solidity of the whole is that of its weakest link. The grid of the three acts gives this requirement its name and its reason.

Inscription Founds; Exchange Extends

It remains to measure the reach of the symmetry set out in the preamble. The act of inscription founds the datum: without it, there is only flow, nothing arrested. The act of exchange extends its availability and its possibility of use: without it, the datum exists — constituted, whole — but remains solitary, an orphan inscription that feeds no one. A datum inscribed and never transported is like the water of a canal dug without being connected to any field: it flows in its bed, without fertilizing what would need it.

Here, a singular property of the datum must be called upon, for it changes the very nature of exchange. When one transmits a material object, one is dispossessed of it: it leaves; one no longer has it. The datum does not work like that — The Two Times of the Datum established it: it is not in its medium but carried by it, and it is not ceded like a thing; it is shared like a fact — the one who transmits keeps a copy, without anyone being dispossessed of it. What circulates, in the act of exchange, is therefore not an object passing from hand to hand; it is a finding that becomes present in several places at once. This is what distinguishes the act of exchange from any transfer of an object; the mechanics of this property — what is counted when one copies, what remains one — is the business of The Properties of Movement.

This non-rivalry is ontological: it lies in what the datum is, not in what the law provides about it. It says nothing about legal regimes — those of the medium, the database, the secret, the person, access and its control — which remain whole and distinct: that sharing does not dispossess does not mean that every sharing is permitted, nor that nothing belongs to anyone. The tome develops none of this; it was enough that property not be confused with a thesis on belonging.

One can then put forward, cautiously, a moderate thesis. When humans do not merely keep their inscriptions but share them, circulation extends the availability and the possibility of use of the datum, without accomplishing either the datum or reading: a frequent and structuring regime, not a universal destination. This is what the etymology indicated, what Whitehead said, what the history of the techniques of inscription seems to show — from the first notches, which counted as shareable markers, to the contemporary flow that addresses countless recipients.

But “being made for circulation” does not mean “circulating well”. Transport has its forms, its constraints, its properties, which can serve the recipient or deceive them, reach them or miss them. That is the subject of the next document, which will examine what, precisely, a transport is — and will discover that it is not what one believes: not the movement of the datum, but a relation between two movements.

Landmarks

For the references used in this document — the Maussian triad “to give, to receive, to return”, the etymological clue of the datum in Whitehead —, see the Annotated Bibliography of Tome 2. The methodological status is set out in the Preface to the Edition; the scroll, the letter, the fibre, the canal without a field belong to the status of illustrations. The concepts taken up from the first tome refer back to their loci: the act of inscription and fixation, as well as the datum carried and not ceded — sharing without dispossession —, to The Two Times of the Datum; the caused trace and the intended datum, with the responsibility of the drawing, to The Trace and the Reader; the flow of observation, to Digging the Canal. The moderate thesis of exchange — circulation, a frequent and structuring regime of the datum, not its universal destination — is an established result of the first edition, which this edition credits and scopes.

This document carries four loci for the trilogy. The act of exchange is established here — a second gesture, human, intentional, situated, dated and unrepeatable, carrying a cost and a responsibility of its own, that of putting into circulation. The transport medium receives its minimal typology here — media that are moved, those that transmit, those that make accessible. The grid of the three acts is closed here — inscription, exchange, reading — and, on the side of reading, not a third medium but the conditions of interpretive resumption, of which this document is the locus. The verb of exchange is fixed here — exchange extends availability and the possibility of use — and non-rivalry receives its scope here: ontological, distinct from legal regimes. Everywhere else, these established results circulate without being demonstrated again.

Transport and Propagation

The Relation between Two Movements, and What Circulation Carries Without Giving It

Tome 2 — Document 3

“Knowledge is the only thing that grows when it is shared.”

— Attributed to Socrates

Preamble — A Naive View to Be Set Right

The previous document set out the act of exchange and the transport medium. It remains to examine what transport itself precisely is. And here a naive view first imposes itself — one to be set right, for everything else depends on it.

The naive view is this: transporting a datum would mean moving it, the datum itself, from one point to another — as one moves a parcel. The sender pushes the datum towards the recipient, and transport is this movement of the datum. This image has a share of truth, but it is incomplete, and its incompleteness masks the essential.

For the recipient is not a motionless point that would only have to receive. They too, often, must move — towards the datum, towards the place where it is waiting for them, towards the shelf where it is kept. Transport is therefore not the movement of the datum alone: it is a relation between two movements, that of the datum towards the reader, and that of the reader towards the datum. This is the thesis of this document, and it changes the way all circulation is thought.

This document finally fixes a vocabulary of two degrees. Transport designates the elementary phase — the conveying of a datum towards a reader. Propagation designates the aggregate object — the set of accumulated transports through which a datum reaches a multitude, step by step. The relation of transport to propagation is that of the act of inscription to the flow of observation: the elementary and the aggregate.

Part I — Transport Is a Relation between Two Movements

The Two Halves of the Path

Reading becomes possible only when the datum and its reader meet. Now this coming together can be accomplished from both sides, and the relation between the two can vary enormously.

With a letter, the sender covers most of the path: they address the datum to a targeted recipient, who has little more to do than receive it. The movement is almost entirely on the side of the datum. With the library, the division is the reverse: the datum comes as far as the place, stops there, and it is the reader who does the rest — they come, they look for the shelf, they open the book. The movement tips over to the side of the reader. With open publication or mass broadcasting, the movement on the side of the datum is maximal and undifferentiated — the datum is within everyone’s reach, with no targeted recipient — and the whole weight of the relation tips over onto the reader, who must accomplish their whole share: connect, search, come.

There are therefore not two opposed modes of transport — one in which the datum would come, the other in which the reader would go. There is a single relation, whose weight is distributed variously between the two movements. Thinking of transport as a relation, and not as a single movement, gives one the means to understand situations that the naive view confuses.

What Each Side Presupposes

This relation is not symmetrical in its requirements, and the asymmetry is instructive.

When the datum covers most of the path, the sender must know who the recipient is, where they are, when to serve them. Transport then presupposes prior knowledge — an address book, a directory, the intention of reaching a given receiver. It is a personalized relation, which cannot address everyone indiscriminately without becoming prohibitive.

When the reader covers most of the path, the sender does not need to know who will come. They set up a place of access, keep its entrance, ensure its preservation. But then it is the reader who must know where to go: knowledge must be built in the other direction — no longer the directory of recipients, but the catalogue of places of access, the shop signs, knowledge of libraries and repositories. It is a collective relation, which can serve a multitude without changing its nature.

Neither of the two distributions is better in itself. They serve different intentions, and the same corpus can be both made available to whoever knows of it and actively addressed to those one wants to reach. But their distinction belongs to substance: it distributes the burden of transport differently, and it calls for different knowledge upstream.

The Criterion of Accomplishment: The Reader’s Share Made Practicable

There remains a question that the naive view cannot settle: when is a transport accomplished? If one said “when the recipient has received”, one would make the accomplishment of transport depend on an act — reception, reading — that does not belong to transport. Now an answer is needed that does not encroach on reading, since reading is the business of the next tome.

The answer lies in the relation itself. Transport is accomplished when the datum has done its share and the reader’s share has become practicable — that is, when the datum is within reach, accessible, whether the reader reads it or not. What follows — do they come? do they have the code? do they read? — is no longer transport; it is reading. A datum lost on the way has not done its share: transport has remained potential. A datum made available has done its own: transport is accomplished, and it is up to the reader to play.

“Within reach” is therefore not a property of the datum alone: it is a property of the relation. The reader’s share must be not only permitted but practicable — they must have the receiver, the access, the ordered place where things can be found. The broadest broadcasting transports nothing towards someone who cannot connect. Making a datum “available to all” on a medium that no one knows how to reach is having accomplished no transport, whatever the scope of the gesture of emission.

This criterion meets and extends a case that The Two Times of the Datum had isolated: the inaccessible datum — “perfectly alive in itself, and nevertheless lost for a situated reader”. One now sees what this inaccessibility is, from the point of view of movement: a transport whose reader’s share could never be made practicable. The datum may well be emitted and durable; the relation has not been accomplished, because the movement on the reader’s side is impossible. Inaccessibility and the non-accomplishment of transport are one and the same thing, seen from two sides: extinction for the situated reader, and the failure of the relation of transport.

This criterion also settles the case of the “recipient who does not receive” without touching reading. The datum made available and never consulted has been transported — its share is done; the reader’s share is practicable —, even if no one comes. What has not taken place is not transport; it is reading. And reading, here, is only named as the limit at which transport stops.

Part II — From the Elementary to the Aggregate: Propagation

The Datum Propagates from Reader to Reader

A transport leads a datum to a reader. But transports accumulate: the datum that has reached one reader is retransmitted to others, who retransmit it in their turn. From this accumulation is born an aggregate object that the present document calls propagation: the set of points a datum passes through as it reaches, step by step, a multitude.

The relation of propagation to transport reproduces that of the flow of observation to the act of inscription, established in Digging the Canal: the act is elementary; the flow is the aggregate of repeated inscriptions. Likewise, transport is elementary; propagation is the aggregate of accumulated transports — The Two Times of the Datum had announced it when fixing the four verbs of duplication. This correspondence is not a convenient analogy: it is the same structure — the elementary and the aggregate — found on both sides of the life of the datum, that of its making and that of its circulation.

What Propagation Carries

Propagation is the material vehicle through which, on the collective scale, not only the datum spreads but what humans draw from it. Each reader, in receiving the datum, deciphers it with their corpus, extracts information from it, integrates it into their knowledge — according to the extended chain set out in Digging the Canal. Propagation circulates the material from which, reader after reader, this chain is reconstituted.

This is the deep meaning of the historical movement that From the Notch to the Algorithm described: what was privilege becomes, with time, common — libraries were the example, at first reserved for an elite, then open to all. The public library generalizes the relation of transport favourable to the reader: it lowers the cost of the reader’s share, until anyone can accomplish their half of the movement. Making common is this: not multiplying the datum, but making practicable, for the greatest number, the share that falls to the reader.

What Propagation Does Not Carry

Here a limit must be stated with the greatest clarity, for it is crucial and it reaches out to the next tome: to propagate the datum is not to propagate knowledge.

The datum arrives within reach — accessible, the tome of the reader will say. Information, knowledge, wisdom do not travel with the signal: actualization is not transported. They require the reader to accomplish their share — the movement towards the datum, its deciphering, its integration. Knowledge is not contained in the transported datum; it is produced by a reader who does their share. Circulating the datum accomplishes only half of the relation; the other half is not of the order of transport, and no improvement of transport will be able to accomplish it in the reader’s place.

That is why a society can be crossed by data as never before and be none the more knowledgeable for it. The torrent — we shall come to it — circulates data at a considerable throughput; it does not thereby circulate knowledge in due proportion, for the share that falls to the reader is not transported. Circulation accomplishes what is within its remit, and nothing more. What it leaves intact — the reader’s share — will be the proper subject of the next tome. The present tome confines itself to naming it as its limit: there is, at the end of transport, something that transport does not do.

Part III — The Impediment of the Relation

A Relation Constrained in Two Ways

The relation of transport is not systematic. It is relative to the will, the capacity and the means that humans give themselves. It can be impeded — and its impediment is of two natures, whose distinction belongs to substance.

Voluntary impediment is the secret. An entity closes access in its own interest: ownership of a body of knowledge, a plan kept hidden, the advantage of those who know over those who do not. The secret is a permanent dimension of human life; there will always be a need for secrets. It is not an imperfection to be corrected: it is a choice, sometimes legitimate. The relation is here deliberately broken.

Involuntary impediment is incapacity. No receiver for the wave, no ordered place to find again what has been accumulated, no means of transmitting. The relation is broken by constraint, not by will. And this impediment calls for a response: humans look for ways round, use their knowledge to open a path between the datum and the reader — learning, teaching, technique.

The Work of Opening Paths

This involuntary impediment points, by what it lacks, to a work of its own: that of opening the paths that incapacity closes. Where the relation is broken for want of means, this work restores the practicability of the reader’s share — it builds the receiver, orders the repository, establishes the bridge between the datum and whoever could read it. And it has an exact scope, which needs to be named: it serves the practicability of reading; it remedies incapacity. It is not its vocation to force secrets. The secret belongs to something other than a technical remedy, for it is not a blind constraint but a will — and one does not resolve a will as one resolves an incapacity.

This distinction runs parallel to the founding one of The Trace and the Reader: the trace is caused; the datum is intended. Involuntary impediment is caused — a blind constraint, remediable; voluntary impediment is intended — an intention to close, which does not fall within the scope of a remedy. The legitimacy of the work of opening paths lies in this divide: it works on the side of the caused, where a path is lacking, not on the side of the intended, where a door has been closed on purpose. This work of opening paths — which opens the way to a data craft — will be treated on its own, outside the present tome.

A Part Deliberately Left Open

Part of this question remains open, and its very openness is significant. Which constraints weigh at a given era, which secrets a society keeps, which means it gives itself to open paths — this the essay cannot settle in its reader’s place. Substance sets the framework: transport is a constrained relation, impeded in two ways, one of which calls for a remedy and the other not. The actual settling — where to place the divide between the legitimate secret and the incapacity to be lifted, which paths to open first — belongs to each person’s present: it is an opening towards form.

Part IV — What Transport and Propagation Illuminate

Irrigation

The essay’s network of images set out the river — the flow of the real —, the canal — the flow of observation —, the spring — the stable point where the real flows slowly, planted by humans between their two thresholds of perception. This tome has added the spiral to it; circulation invites it to add a second image: irrigation.

Irrigation is the art of bringing water from where it flows to where it is needed — the field, the garden, the rice paddy. Without irrigation, the canal may well be dug, but the water would stay in its bed, without fertilizing what needs it. Irrigation is the completion of the canal: what makes it useful, that without which the digging would have been in vain. The same holds for the datum: without transport, inscription is solitary. The flow of observation may well be well built, its optimum well located, its probity high — if it feeds no one, it remains without effect. Transport is what irrigates society with the data it produces.

This image also states the political dimension of transport. Irrigation is not neutral: it has its beneficiaries and those it leaves out. An irrigation policy decides who will be watered and who will stay dry. The same holds for transport: who benefits from it, who bears its cost, who decides its flows — these are political questions, which arise as soon as there is an asymmetry between the producers of inscriptions and the possible recipients. The political problem of the datum begins with inscription; it is completed — or it worsens — with transport.

Contact as the Essence of Transport

One last clarification brings out the essential. Since transport is a relation between two movements, what defines it is not who moves towards whom — the datum towards the reader, or the reverse — but the contact that results: the moment when each one’s share has been done and the meeting has become practicable. Whether the datum comes or the reader goes is secondary; what is primary is that the two halves of the path meet and that a contact becomes possible. Transport is the establishment of this contact.

And it is there that it must be left for the rest of the tome. Once it is established that transport establishes a practicable contact, one can examine the properties of this movement — its fidelity, its duration, its singular capacity to multiply what it transports —, then see how, by accumulating and accelerating, propagation engenders the regime that the essay has named the torrent. The rest of the tome follows this thread. But the crucial point established here will not be forgotten: at the end of the best-established contact, there remains the reader’s share, which transport makes practicable without ever accomplishing it.

Landmarks

For the references used in this document — the “privilege → common” movement of libraries, Elinor Ostrom’s theory of the commons in the background of the political dimension of transport —, see the Annotated Bibliography of Tome 2. The methodological status is set out in the Preface to the Edition; the letter, the library, mass broadcasting, irrigation belong to the status of illustrations and of guiding images. The concepts taken up refer back to their loci: the extended chain and the flow of observation, with the relation of the elementary to the aggregate, to Digging the Canal; the inaccessible datum and the four verbs of duplication, to The Two Times of the Datum; the caused trace and the intended datum, to The Trace and the Reader; the history of libraries, to From the Notch to the Algorithm.

This document carries, for the trilogy, five loci. Transport as a relation between two movements is demonstrated here — the datum towards the reader, the reader towards the datum — with its criterion of accomplishment (the datum’s share done, the reader’s share made practicable) and its essence: the establishment of a contact. The vocabulary of two degrees is fixed here — transport, elementary; propagation, the aggregate of accumulated transports. The thesis “to propagate the datum is not to propagate knowledge” receives its demonstration here: the reader’s share is not transported, and no improvement of transport will accomplish it in the reader’s place. The impediment of the relation is divided here — voluntary, the secret; involuntary, incapacity — and the work of opening paths receives its scope here: on the side of the caused, not of the intended. The image of irrigation, finally, enters here into the network of guiding images, with the political dimension of transport. Everywhere else, these established results circulate without being demonstrated again.

The Properties of Movement

Fidelity, Duration, and the Strange Power of Multiplying the Carriers of What One Transports

Tome 2 — Document 4

“He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me.”

— Thomas Jefferson, letter to Isaac McPherson (1813)

Preamble — Transport Is Not Neutral

The previous document established what transport is: a relation between two movements, accomplished when the datum has done its share and the reader’s share is practicable. But saying that a transport has taken place does not yet say how it took place. A transport can be faithful or betraying, fast or slow, conserving or multiplying. Like inscription, transport has its properties, and these properties are not neutral: they decide what arrives at the end of the path.

Digging the Canal had listed the properties of the flow of observation; the present document lists, in the same spirit, those of transport. Three hold our attention, because they govern the whole rest of the tome: transport has its fidelity, it has its temporality, and it multiplies. The third is the most singular, and it is through it that the movement of the datum parts company most clearly with the movement of objects.

Part I — The Fidelity of Transport

A Fidelity Distinct from That of Inscription

A transport can be more or less faithful. A copy can be exact or approximate, a translation accurate or betraying, a transcription rigorous or faulty. At each relay — each intermediate medium, each conversion, each recopying —, the datum runs the risk of alteration.

This fidelity of transport is distinct from the fidelity of inscription, which Digging the Canal had set out as one of the variables of the observational optimum, under the name of probity. Inscription can be perfectly faithful to its flow of observation, and transport then degrade this inscription: the accurate datum, badly transported, arrives distorted. Conversely, an approximate inscription can be transported identically, without transport adding any fault to it: the fidelity of transport does not correct the infidelity of inscription; it only does not aggravate it. The two fidelities are independent, and they must be judged separately. Asking whether a datum is faithful is in reality asking two questions: to what was it faithfully drawn, and was it faithfully conveyed.

The Five Planes of Fidelity

But “faithfully conveyed” remains too short an answer, for the fidelity of transport is not a single quantity: it is judged plane by plane. At each relay, the guiding question is the same: what is invariant? And it arises on five planes. The physical plane — the material state of the signal: does the medium hold, does the wave still carry. The syntactic plane — the sequence of signs: the characters, the bits, what an exact copy reproduces identically. The structural plane — the organization: the order of the parts, the links, the schema that articulates them. The semantic plane — what the datum states: the proposition, what an accurate translation preserves while changing all the signs. The pragmatic plane — what the datum makes it possible to do: its aptitude for use, which can be lost even while all the rest holds.

These five planes do not vary together, and this is what makes the question fruitful. A digital copy is invariant up to the syntactic plane and can nevertheless lose the pragmatic plane — the intact file that no interface any longer knows how to open. A translation overturns the syntactic plane and preserves, if it is accurate, the semantic plane — that is even its definition. A condensation sacrifices some structure to keep the essence of the meaning. To say of a transport that it is faithful without saying on which plane is to say nothing precise: the exact or approximate copy, the accurate or betraying translation, the rigorous or faulty transcription — all the cases that open this part are settled by the question of the invariant.

The Transport Optimum

There is no doubt a transport optimum, on the model of the observational optimum established in Digging the Canal — which already counted transport among the costs to be arbitrated. Complete fidelity is not always the aim: there is a point where fidelity is sufficient for the use — where the planes the use requires are preserved, without demanding the invariance of all of them where it serves no purpose.

A shorthand transcription of a technical speech requires great fidelity, for every term counts. The account of a discovery intended for the general public can be very condensed without harming its object — provided it is honest about its own condensation. Good transport is therefore not complete transport: it is the transport that serves its recipient without deceiving them. Transporting everything, always, at the highest resolution, would be to the fidelity of transport what over-engineering was to observation: a cost incurred without benefit, sometimes to the detriment of clarity.

The probity of transport lies precisely there: not in the impossible promise of absolute fidelity, but in honesty about the part that has been condensed. An acknowledged condensation serves; a masked condensation deceives. It is the version, on the side of movement, of the probity of drawing that Digging the Canal required of inscription.

Part II — The Temporality of Transport

Every Transport Takes Time

Every transport takes time. This time can be negligible — a signal between two neighbouring components — or considerable — a manuscript travelling for months. But it is never zero. No transport is instantaneous in the strict sense; the fastest still has a duration, however imperceptible.

Now, while the datum travels, the flow of the real continues to flow. The Datum Navigates Because It Is Fixed established it: the two navigations run together, and the trip adds its own ageing to the ageing of origin — transport is itself a cause of lapse. What the present document adds concerns duration itself: as long as it is perceptible, it can be seen; it is when it disappears that it acts most.

The Disappearance of Delay and the Warning It Carried

Digging the Canal had identified a systemic leap in the instantaneity of transmission, and The Datum Navigates Because It Is Fixed measured its nature: the trip’s own duration fallen below the threshold of perception, and with it transport ageing. Here is what this disappearance does, which acceleration alone does not explain.

As long as transport had a perceptible duration, everyone knew, on receiving a datum, that it had aged on the way — the letter told news that was already old, and this was taken into account. Delay was not only a nuisance: it was a warning, carried by the very form of transport. Near-instantaneous transmission erases delay and, with it, the warning. The datum seems to arrive from the present itself — whereas abolishing transport ageing does not abolish ageing of origin: an old datum transmitted instantaneously remains old. Instantaneity makes disappear the only delay it could make disappear, and thereby masks the delay it does not touch: transport freshness can pass itself off as a freshness of origin that is not there.

Part III — Transport Multiplies

Transporting a Datum Does Not Move It: It Multiplies Its Carriers

Here is the most singular property of the datum with regard to material objects. When one transports an object, it leaves its place of origin: it is here or it is there, never in both. When one transports a datum — in the duplicative mode, the dominant case for the digital datum and already, in part, for the written datum copied out or printed, unlike the movement of a carrier medium, letter or tablet, in which the same occurrence changes place —, the original inscription stays where it was. Transport, in this case, does not move: it multiplies.

This property links up with the rule that The Two Times of the Datum fixed in four verbs: the act is preserved, the inscription is duplicated, the carriers multiply, the datum is shared. This document keeps the promise that accompanied this rule: to give the structure its names, and to establish its mechanics — what is counted when one copies, what remains one.

What Is Counted and What Remains: The Occurrence and the Datum of Origin

What is counted, when one copies, are occurrences: the singular inscription of a datum on a medium, in a place, at an instant. The manuscript on its shelf is an occurrence; its copy on the next shelf is a second; the file on each machine, one more. Transmitting multiplies occurrences — and an occurrence, once it has taken place, is not crossed out: revision does not erase the inscription; it produces a new one.

What remains one is the datum of origin: what the act of inscription brought into being and what every copy instantiates. It is identified by its origin — which act, on which fragment of the real, under which criteria — and not by the number of its carriers. Occurrences multiply; the datum of origin remains one. This is copying without division, now seen from both sides: on the side of the medium, inscriptions that are counted; on the side of the act, a dated drawing that has no more reason to split in two when it is copied than an event has to happen again when it is told twice.

The familiar formula “the two copies are the same datum” receives its precision here. It is defensible if it means: two occurrences realize the same datum of origin. It becomes false if it claims that there is only one material object — there are indeed two inscriptions, on two media, which any inventory would count. And the criterion works in both directions: two recordings carrying the same value can be two data, if they were born of two distinct acts — two sensors measuring the same temperature at the same instant produce two drawings, not one; two different files can realize the same datum, if they are faithful copies of the same inscription. The identity of the datum is read neither in the value nor in the medium: it is read in the origin.

Copy versus Filiation

Not everything that circulates is a copy, however. Correcting a faulty value, translating a text, aggregating a thousand entries into an average, summarizing a report, inferring an estimate: these operations produce, from a parent datum, a new datum — a descendant, not a copy. The copy multiplies the occurrences of one datum; transformation opens a filiation: a line of descent of data, each born of the previous one through an operation that has changed something — the content, the status, the object aimed at, the validity regime.

The boundary between the two is not natural, and this is a point this tome must hand on with care: it is defined by the invariants one intends to preserve. The five planes of fidelity give the question that settles it. A conversion that preserves the planes the use requires is still copying — faithful on what counts, displaced on the rest; an operation that alters one of these planes — meaning condensed, structure recast, aptitude changed — engenders a descendant. Translation is the textbook case: it overturns all the signs and preserves, if it is accurate, the proposition; copy or descendant, this depends on the invariants a practice declares it holds. In one line: the faithful copy maintains, conversion displaces, derivation engenders. The derived datum, whose genus The Flow and the Annotation set out, receives its mechanics here: every derived datum opens a filiation — and only the operation that alters a declared invariant is a derivation, faithful conversion producing only a new occurrence of the same datum.

Provenance, Constitutive

It remains to name what holds the whole together. If the identity of the datum is read in the origin, and if filiation is a line of descent, then the datum must be able to say how it was made: from which act it was born, through which operations and which hands it has passed. This is its provenance. It is not an ornament of erudition that would be added to important data: it is constitutive. Without it, identity loses its footing — there is no longer any way of saying whether two occurrences realize the same datum, whether a given recording is a copy or a descendant — and imputability loses its footing with it: there is no longer any way of saying who answers for what. For a derived datum, provenance includes the model and the parent data: saying where an average comes from means saying from which entries, and by which calculation.

The identity of the datum is therefore stratified, and its strata are historical: an act of origin, which makes the datum; occurrences, which are counted; a filiation, when transformation engenders; a provenance, which documents the whole. The copy multiplies occurrences, transformation engenders a filiation, provenance holds the whole together — here, named piece by piece, are the mechanics of the sharing without dispossession that The Act of Exchange announced.

Why This Property Baffles Inherited Frameworks

This property is not a detail. It is what makes the datum so difficult to think within the frameworks inherited from property law over objects. A material theft deprives the owner of their good; a copy deprives them only of exclusivity — and one now sees why, with the names of this document: what one holds is an occurrence; the datum, identified by its act and not by its carriers, is not the kind of thing that possession exhausts. The Flow and the Annotation had sensed it in distinguishing the ownership of the drawing from the inappropriability of the flow, and set out the scope that suffices here: non-rivalry is ontological; legal regimes are discussed in their own order. A framework designed for objects that move without multiplying will always be at odds with an entity that is shared without being ceded.

The Three Axes of Duplication

Duplication, which thus lies at the heart of transport, can be characterized along three axes, which measure its qualities and make it possible to compare its forms.

Cost: what the copy requires in means, energy, production time. It has decreased over the course of techniques, until it has become, for the digital datum, so low that it is neglected — wrongly, for it is never zero.

Speed: the rapidity with which a copy is produced and conveyed. It governs the transport ageing examined above: the faster the duplication, the less the copy has aged on the way.

Fidelity: the degree to which the copy reproduces the original — plane by plane, as the first part established. Digital duplication reaches invariance down to the sign — the copy is strictly identical —, where manual copying introduced deviations at each relay.

These three axes do not vary together. A copy can be faithful but slow, fast but costly, cheap but unfaithful. Characterizing a form of transport means situating its duplication on these three axes. And it is their conjunction — near-zero cost, near-instantaneous speed, perfect fidelity — that defines the contemporary regime of duplication, and prepares the ground for the torrent.

Part IV — What These Properties Announce

Potentially Unlimited Diffusion

From multiplication follows a consequence that the following documents will unfold. Where a material transport is an act of finite movement — a good goes from a single point to a single point —, the transport of a datum by copying opens up the indefiniteness of the number of recipients. Once transmitted a first time, the datum can be transmitted again, and again, without proportionate marginal cost. The transport of the datum is, by construction, an act of potential diffusion.

It is this property that makes great diffusions and open commons possible; it is also what makes leaks and abusive uses hard to contain. For better and for worse, the transported datum is a datum with potentially multiplied carriers. The gesture of exchange, which decided a transport, in reality engages a diffusion whose limits it does not control — and one measures here the full reach of the responsibility of putting into circulation, which The Act of Exchange established.

From Movement to Its Apparent Stability

The properties listed here — variable fidelity, transport ageing, multiplication — describe the movement of the datum from one place to another. But there is a case that these properties illuminate by contrast: that of data that seem not to move, stable sources, the references one consults without their seeming to age. How can a datum seem motionless when the real never stops turning around it? This is the subject of the next document: navigating in place — apparent stability as a certain ratio between the speed of use and the speed of lapse.

Landmarks

For the references used in this document — Jefferson’s letter to McPherson in the epigraph, the datum’s property of duplication, the distinction between the ownership of the drawing and the inappropriability of the flow —, see the Annotated Bibliography of Tome 2. The methodological status is set out in the Preface to the Edition; shorthand, the press report, the manuscript, the digital copy belong to the status of illustrations. The concepts taken up refer back to their loci: the observational optimum, the probity of drawing, the systemic leap and the properties of the flow of observation, to Digging the Canal; the four verbs of duplication and copying without division, to The Two Times of the Datum; the two navigations and the divide between the two ageings, to The Datum Navigates Because It Is Fixed; the genus of the derived datum and the scope of non-rivalry, to The Flow and the Annotation; the responsibility of putting into circulation, to The Act of Exchange.

This document carries, for the trilogy, the locus of stratified identity. The datum of origin and the occurrence receive their names here — what remains one, what is counted — with the precision of “the two copies are the same datum” and revision as a new inscription. Copy versus filiation is divided here — the faithful copy maintains, conversion displaces, derivation engenders — the boundary being defined by the invariants a practice intends to preserve; the derived datum receives its mechanics here. Provenance is established here as constitutive: without it, identity and imputability lose their footing; for a derived datum, it includes the model and the parent data. The five planes of fidelity are fixed here — physical, syntactic, structural, semantic, pragmatic, under the guiding question of the invariant — with the transport optimum and the probity of acknowledged condensation. Temporality yields here the document’s own contribution — perceptible delay carried a warning, which instantaneity erases — and the three axes of duplication (cost, speed, fidelity), with potential diffusion, prepare the documents that follow. Everywhere else, these established results circulate without being demonstrated again.

Navigating in Place

Stable Sources, or the Datum That Seems Motionless Because the Real Flows Slowly

Tome 2 — Document 5

Preamble — The Illusion of the Motionless

The Datum Navigates Because It Is Fixed set out double navigation — the journey in space, carried by a medium; the real moving away while turning around the marked point — then the following documents examined the triggering of movement, its structure, its properties. There remains a case that seems to escape all navigation: that of data one consults as reliable points of support, and which seem not to move. The table of an established body of knowledge, a physical constant, the map of a territory, a dictionary: one consults them today as yesterday, and they seem to say the same thing. They seem motionless.

This immobility is an illusion, and dispelling it sheds light in return on the whole movement of the datum. No datum escapes temporal navigation; The Two Times of the Datum established it without exception: from the instant of inscription, the real begins to move away from the marked point. If certain data seem stable, it is not that the real does not move away from them — it is that it moves away from them so slowly, relative to the use one makes of them, that the drift is imperceptible on the scale of that use. Stability is not the absence of movement; it is a certain ratio between two speeds.

Part I — Stability as a Ratio of Speeds

Two Speeds at Play

The first speed is an established result: lapse is a relation between a motionless datum and a mobile real, and its speed is not a property of the datum — it is that of the fragment of the real from which it was extracted (The Two Times of the Datum).

To this first speed — that of lapse, imposed by the real — the present document adds a second: the speed of use, the frequency and the time scale at which the datum is used. One consults a dictionary several times a day, but the meaning of words drifts over decades. One consults a physical constant constantly, but its established value does not move on the scale of a lifetime. Apparent stability arises from the meeting of these two speeds: it is a certain ratio between the speed of use and the speed of lapse — a datum seems stable when its speed of lapse is far lower than its speed of use, when the real it draws moves away much more slowly than one uses it.

Why the Same Datum Is Stable Here, Fleeting There

This ratio of speeds explains why the same datum can be held to be stable in one use and fleeting in another. The position of a coastline is a stable datum for the walker who finds their way from one week to the next; it is a fleeting datum for the geomorphologist who studies erosion over the century. Nothing has changed in the datum: it is the scale of use that has changed, and with it the ratio between the two speeds. Stability is therefore never absolute; it is always relative to a use — in the very sense in which validity is validity for a use (The Two Times of the Datum) — and it is the use that sets the scale at which one looks.

That is why speaking of “stable data” in themselves is as misleading as speaking of “raw data” — the oxymoron that The Flow and the Annotation undid: it attributes to the datum a property that in reality belongs to the relation between the datum and the use one makes of it. The Two Times of the Datum established the stable source — that fragment of the real whose lapse is so slow that it can be held, on the scale of use, to be negligible — and it is there that humans, between their threshold of the instantaneous and their threshold of the eternal, plant their springs. The present document gives only its translation into speeds: a stable source is a source whose real moves away slowly enough for the use intended for it, not a source withdrawn from the flowing; and planting a spring means choosing, in the flow, a point whose drift is slow enough to lean on.

Part II — Navigating in Place

The Movement of the One Who Consults

One can then understand what “navigating in place” means. When one comes back to consult a stable source, one does not move it — it stays where it is, and Transport and Propagation established it: a datum made available has done its share of transport. It is we who move towards it, again and again. The stable source is the limit case of the relation between the two movements in which all the weight is on the reader’s side: the datum does not move; it is the reader who, at each use, travels the path to it again.

But while the reader navigates in this way towards the source, the real goes on turning around the datum. The reader comes back to the same point; the point, imperceptibly, is no longer quite at the same place in the flow. “Navigating in place” is this: coming back to a datum one believes motionless, while the real has drifted slowly around it. The movement is not in the datum one consults, nor in the reader who comes back to it; it is in the gap, which widens quietly, between what the datum said at its inscription and what the present real says.

The Danger Proper to Stable Sources

From this comes a danger proper to these sources, which is the reverse of their virtue. Their virtue is that one can lean on them without fearing a rapid drift. Their danger is that one ends up forgetting that they drift. A manifestly fleeting datum carries on it the warning of its freshness: no one decides about the present with the stock market price of ten years ago. But a datum held to be stable lends itself to the forgetting of its own temporal navigation: one consults it for years as if it still told the present, until the day when the accumulated drift makes it misleading without anything, on its surface, having signalled it.

The stale datum is then treated as fresh. Stable sources are its favourite ground, precisely because their drift is slow and silent. The reader’s probity — a limit of this tome, which must nevertheless be named — would consist in knowing, each time one leans on a stable source, how long ago the real left the point it marks: the vigilance that The Two Times of the Datum asked for every datum holds here doubly, because nothing, on the surface of a stable source, recalls it. The tome only sets out the ratio of speeds that makes this vigilance necessary; what the reader does with it belongs to them.

Part III — What Stable Sources Illuminate

There Is No Outside-the-Flow

The lesson of substance is that there is no outside-the-flow. The first tome had fought the idea that the datum would be an object withdrawn from time; the present document fights its last refuge: the idea that there would be, at least, certain data — the most established, the most certain — that had left the current. There are none. The most stable sources drift like the others; the real simply moves away from them more slowly. Their stability is a regime of navigation, not an exception to it.

This absence of an outside-the-flow prepares the central concept of the end of the tome. For if even the slowest sources drift, then the general acceleration of the production of data leaves nothing intact: it does not only multiply fleeting data; it also accelerates, by contagion, the real that stable sources drew slowly. When the real itself is set in motion by human action — what the tome will soon call the “real-marked-by-us” regime —, the sources that were stable because their real moved away slowly lose part of their stability: their real now moves away faster. Stability, which was a ratio of speeds, degrades when one of the two speeds races.

Towards the Torrent

The tome has now established everything needed to think the torrent. It has set out that the datum navigates because it is fixed, that an act of exchange triggers its transport, that transport is a relation between two movements of which propagation is the aggregate, that this movement has its properties — among them multiplication —, and that no datum, not even the most stable, escapes temporal navigation. It remains to see what happens when this movement, in all its dimensions, accelerates to the point of changing regime. This is the moment that the first tome called the torrent (The Wave and the Ocean), and it is the subject of the last three documents of the tome.

Landmarks

This document uses no external reference of its own; the tome’s apparatus remains in the Annotated Bibliography of Tome 2. The methodological status is set out in the Preface to the Edition; the dictionary, the physical constant, the map of a territory, the coastline seen by the walker and by the geomorphologist belong to the status of illustrations. The concepts taken up refer back to their loci: lapse as a relation between a motionless datum and a mobile real, the speed of the fragment of the real, the stable source and the two thresholds between which humans plant their springs, as well as validity for a use, to The Two Times of the Datum; the oxymoron of “raw data”, to The Flow and the Annotation; transport as a relation between two movements and the datum made available that has done its share, to Transport and Propagation; the shoal of the stale datum treated as fresh, to The Wave and the Ocean; double navigation and the spiral of the real around the marked point, to The Datum Navigates Because It Is Fixed.

This document carries, for the trilogy, the locus of apparent stability: a certain ratio between the speed of use and the speed of lapse — the stable source translated into speeds, never withdrawn from the flowing. Navigating in place receives its meaning here: the reader travels the path again towards a motionless datum while the real drifts around it, and the movement is in the gap, which widens quietly. The degradation of stability is established here: when human action accelerates the real, one of the two speeds races and stable sources lose part of their footing — the bridge to the torrent. Everywhere else, these established results circulate without being demonstrated again.

The Torrent

Definition of a Regime, and Why It Carries Our Signature

Tome 2 — Document 6

“We are drowning in information but starved for knowledge.”

— John Naisbitt, Megatrends (1982)

Preamble — Taking Up the Torrent Again, but Grounding It

The first tome ended on an image: the torrent — running water, a wealth, a force; and what can carry away, drown, escape (The Wave and the Ocean). Humanity no longer engraves notches; it permanently tends countless flows of observation: the word designated this moment, and the first tome warned that its theory — throughputs, the queue — would be the business of the tome of movement.

This document takes up the image again, but it no longer merely names it: it grounds it on everything the tome has established. The torrent is no longer an opening metaphor; it is now a concept that double navigation, the act of exchange, propagation and the properties of movement make it possible to define precisely. And this definition brings out a point that the image alone left in shadow: the torrent is not a flood imposed from outside. It carries our signature.

Part I — Technical Definition of the Torrent

What Distinguishes the Torrent from the Flow

The flow of the real flows of itself; no one produces it. The flow of observation is built by humans, but each flow taken in isolation remains measured, framed by an intention. The torrent is something else: it is the regime that the circulation of the datum reaches when the multitude of flows of observation, their transport and their propagation accelerate and aggregate to the point that the overall throughput changes nature.

The torrent is defined by the conjunction of several features that the tome has isolated one by one. There is, first, the multiplicity of flows of observation tended permanently — no longer one-off drawings, but countless canals capturing a continuous flowing. There is, next, the speed of transport, pushed to near-instantaneity, which abolishes transport ageing and makes data arrive as if they came out of the present. There is, finally, multiplication by copying, which opens up the indefiniteness of the number of recipients and multiplies the carriers of each datum without proportionate cost. Multiplicity of sources, speed of transport, multiplication by copying: the first of these features is an established result of the first tome (Digging the Canal); the other two, of the present tome (The Datum Navigates Because It Is Fixed; The Properties of Movement, which weighed their conjunction — near-zero cost, near-instantaneous speed, perfect fidelity — and announced the ground of the torrent prepared). Together, they define the torrent.

The Torrent Is Not a Stock

An error must be set aside, the one that The Wave and the Ocean set aside in advance by defining the torrent as a throughput: the torrent is not a stock that accumulates. The datum is not an object but a drawing (The Flow and the Annotation); the torrent is therefore not a reserve that grows, but a multitude of drawings in movement on a real that continues to flow. To believe that the torrent is a stock is to fall back into the object ontology that the whole essay has undone. The torrent does not pile up: it circulates. And what circulates ages while it circulates, multiplies as it propagates, and increasingly exceeds what any window can receive.

That is why storing is not knowing, and accumulating is not understanding. The torrent can grow indefinitely without any knowledge growing in proportion — for to propagate the datum is not to propagate knowledge (Transport and Propagation): it requires the reader’s share, which the torrent does not supply. A torrent is a mass of data in circulation, not a mass of knowledge.

The Mismatch of Cadences

Exceeding what a window can receive: the formula states the fact; it does not yet give its structure. The technical definition of the regime lies in a mismatch. Four cadences run together in every circulation: that of production — what arrives; that of selection — what a rule retains; that of processing — what a capacity of reading can absorb; that of the deadline — the delay beyond which an item loses its value for the intended use. As long as these cadences remain commensurate, circulation absorbs them; the torrent is the regime in which they become durably mismatched. What arrives exceeds what is processed, and the wait lengthens while the deadline approaches: a queue forms — not a reserve that piles up, but a wait that stretches. Its measurement belongs to the formal machinery and lives in the Architecture Volume; prose needs only the structure.

This queue is the exact place where the addition of the two navigations, set out at the opening of the tome, is the whole of the announced problem (The Datum Navigates Because It Is Fixed): while an item waits to be read, the real goes on moving away from it. Waiting is an ageing; the torrent makes age in the queue what it makes arrive fresh.

The Seven Outcomes of the Queue

What becomes of an item that has entered the queue? The torrent decomposed is its outcomes — seven, which prose can name without their formal machinery. Accumulated delay: the item will be read, but later than it should have been. Expiry in the queue: the deadline passes before reading — stale for its use without having been read. Explicit rejection: a rule sets it aside, and the setting aside is documented. Silent loss: the item drops without anything recording it — the only outcome that leaves no provenance. Aggregation: several items reduced to a synthetic grain that sums them up. Prioritization: the order of waiting reordered according to a rule — some go ahead; the others will wait that much longer. Delegation: processing entrusted to another reader, human or automaton, which moves the queue without abolishing it.

These outcomes are not ontologically equivalent, and it is here that the tome’s ontology guards the torrent against false resolutions. A datum in waiting does not become a trace: it remains constituted, simply unread — not actualized, the tome of the reader will say. A rejected datum remains a datum: rejection is a decision of use, not a change of being. A datum expired for one use can remain valid for another — validity is validity for a use (The Two Times of the Datum): stale for the alert, accurate for history. And aggregation is not a loss: it engenders a derived datum, endowed with its filiation and its provenance (The Properties of Movement).

The Five Throughputs

There remains what the mismatch, by itself, does not say. The torrent does not establish the goodness of what arrives; it establishes an incompatibility of cadence. To think it without moralizing it, one must distinguish five throughputs that usage confuses. Raw throughput: everything that arrives. Relevant throughput: what, within the raw, concerns the intended use. Priority throughput: what the rule places at the head of the queue. Processable throughput: what the capacity of reading can absorb within the deadline. Throughput actually read: what a reader actualizes. In the regime of the torrent, these five throughputs never coincide, and the gap between the first and the last is the very measure of the problem.

Between the raw and the read, a rule operates — it selects, orders, aggregates, rejects: it is the rule that transforms raw throughput into throughput presented to the reader. This rule is not a property of the torrent; it is a decision — political in the proper sense: it arbitrates what will be put before eyes and what will wait, be aggregated, expire or drop. Who decides, and by what right: this is what the rest of the document establishes — the torrent carries our signature down to its rules of presentation.

Part II — The Torrent as the Regime of the Real Marked by Us

One Real, Two Regimes

Here is the point that the image alone left in shadow, and which must run through the whole rest of the tome without being diluted. One would be tempted to oppose two flows: the one natural and pure, the flow of the real; the other artificial, the torrent of data. This opposition is false, and correcting it is decisive.

There is only one real that flows. The Wave and the Ocean established it: humans do not observe the flow from an outer bank — they are a portion of it, become capable of reading the others, and the wave is made of the ocean’s water. What humans produce — the torrent — is therefore also flow of the real; it is not another substance.

The right distinction is not real / artefact, but the real-without-us / the real-marked-by-us: one real, two regimes, depending on whether human action is inscribed in it or not. The torrent is not what the flow of the real would have produced if humans had not acted. It is, within the flow of the real, the resultant of human impact — the part of the flow that carries the signature of human action. Distinguishing it counterfactually, by what it would not be without us, means attaching it to the real without naturalizing it.

We Set the Speed

From this follows directly responsibility, and it is the crucial consequence of this document. One cannot say “the real is going too fast” as one would say “it is raining”. Rain is done to us; the speed of the torrent is something we set. The speed of the flow of data — and therefore of the flow of the real that this torrent impacts in return — is the responsibility of humans.

This safeguard prevents the torrent from being naturalized. A natural flood imposes itself from outside, and one does not answer for it; the torrent is dug by our own acts — of inscription, of exchange, of propagation. It has someone responsible: whoever feeds it. This does not mean that an individual commands it; it means that it is our collective resultant, and that one can declare oneself neither its sole author nor its innocent victim. The torrent is ours, not as property, but as responsibility.

This responsibility extends a thread of The Datum Navigates Because It Is Fixed. Temporal navigation, we said there, has its source in a flowing that no one commands — but humans are not thereby passive, since they impact the real. The torrent is precisely the place where this impact becomes massive — or so the tome maintains, by argument, not by measurement: by accelerating the flow of data, humans accelerate the real they mark, and therefore the speed at which this real moves away from each datum drawn. Where we accelerate the real in this way, lapse accelerates with it. This is one more reason not to naturalize the torrent: not only do we produce it, but in producing it we hasten the ageing of what it carries along.

This responsibility has a consequence that the following documents will unfold. If the torrent were natural, the only sensible attitude would be to protect oneself from it as from bad weather. Because it is ours, another attitude becomes possible — and required: answering for it. The next document will show how the torrent digs itself, through a loop each step of which is reasonable and whose sum is not; and the last document of the tome will lead this responsibility up to the threshold where circulation produces more than any window can receive.

Part III — What the Concept of the Torrent Illuminates

The Shoals, Pointed Out and Not Described

The rule is set out in the Preface to the Edition — the concept is a lamp: it points out the shoals; the detailed chart of the reefs falls to the reader — and The Wave and the Ocean gathered together the essay’s shoals. The present document keeps the rule and adds nothing to the list: it shows how the torrent, as a regime, modulates three shoals already pointed out.

The shoal of confusing the datum with the real: the broader and more fluid the torrent, the more it resembles the river, and the more one forgets that it is only a canal of it — multiplied. The shoal of the stale datum treated as fresh: the near-instantaneity of transport masks ageing of origin under transport freshness — the warning of the delay, erased (The Properties of Movement) — and a torrent of transport-fresh data can be a torrent of data already stale at the source (The Datum Navigates Because It Is Fixed): nothing, on their surface, says how long ago the real left them. The shoal of over-engineering: the torrent encourages drawing ever more, far beyond the observational optimum (Digging the Canal), as if mass were valuable in itself. These shoals, the essay illuminates; it leaves to the reader the task of making, in their present, the chart that the concept cannot make for them.

The Torrent Makes the Problem Thinkable

The contribution of the concept is not to judge the torrent — the essay does not say whether it is a good or an evil; that would be a judgement of form, dated. Its contribution is to make the torrent thinkable. A torrent that is thought is already a torrent in which one no longer merely lets oneself be carried away. The tome has forged the instruments that make it possible to look at it clearly: double navigation shows that it ages as it circulates; propagation shows that it does not transport the knowledge it seems to carry along; the queue and the five throughputs show that its problem is one of cadence, and that presentation in it is a decision; the “real-marked-by-us” regime shows that it is our resultant, not a fatality.

It remains to understand why it grows — not by accident, but through a dynamic of its own. For the torrent does not merely exist: it digs itself, it amplifies itself, it feeds its own flood. It is this dynamic, self-amplification, that the next document examines — and it is this dynamic that will reveal that the torrent grows for want, not of knowledge, but of wisdom.

Landmarks

For the references used in this document — Naisbitt’s epigraph on the abundance of information and the hunger for knowledge —, see the Annotated Bibliography of Tome 2. The methodological status is set out in the Preface to the Edition; the torrent, the flood, the wave and the ocean are guiding images belonging to this status. The concepts taken up refer back to their loci: the moment of the torrent and its name, the portion of the flow and the wave made of the ocean’s water, as well as the gathered shoals, to The Wave and the Ocean; the datum as a constituted entity — drawing being its observational route —, not as an object, to The Flow and the Annotation; tended canals and the observational optimum, to Digging the Canal; double navigation and the addition of the two ageings, to The Datum Navigates Because It Is Fixed; the thesis “to propagate the datum is not to propagate knowledge”, to Transport and Propagation; the conjunction of duplication — cost, speed, fidelity — and the filiation of the derived datum, to The Properties of Movement; validity for a use, to The Two Times of the Datum. The formalization of over-throughput — arrival rate, capacity, queue, deadline — lives in the Architecture Volume.

This document carries, for the trilogy, the share of the tome of movement in the loci of the torrent. The technical definition of the regime is established here: three features — multiplicity, speed, multiplication — and a durable mismatch between four cadences — production, selection, processing, deadline. The queue is decomposed here: seven outcomes — accumulated delay, expiry in the queue, explicit rejection, silent loss, aggregation, prioritization, delegation — ontologically non-equivalent. The five throughputs are distinguished here — raw, relevant, priority, processable, actually read — with the rule that transforms the raw into the presented: a decision, political in the proper sense. The real-without-us / real-marked-by-us divide receives its demonstration here — keeping the announcement made from the opening of the tome — and with it responsibility: we set the speed. The share of the tome of the reader — the over-plus — remains to come, at its locus. Everywhere else, these established results circulate without being demonstrated again.

Self-Amplification

The Loop That Digs Its Own Bed, and the Silence of Wisdom

Tome 2 — Document 7

“It is not enough to have a good mind; the main thing is to apply it well.”

— René Descartes, Discourse on the Method (1637)

Preamble — Why the Torrent Grows

The previous document defined the torrent and established that it carries our signature: it is the real marked by us, and we set its speed. But saying that the torrent is ours does not yet say why it grows. One might believe that it grows by decision — that someone, somewhere, wants it to swell. This is not the case, and that is what makes the thing formidable. The torrent grows without anyone having to want its growth as such, through dynamics that produce the conditions of their own continuation.

There are two of them, complementary and of equal rank. The first is a loop: more data call for more automation, which itself produces more data. The second is the coexistence of regimes: when a new regime of observation sets in, the old one does not disappear at once, and the two are added together for the time of a transition that drags on. The loop digs the torrent’s bed; coexistence widens its banks. This document describes these two engines in turn, then proposes their common diagnosis by means of the extended chain set out in Digging the Canal. The diagnosis is precise, and it shifts the location of the problem: neither of these engines arises from a lack of knowledge; both arise from a lack of wisdom. The torrent grows not through an excess of stupidity, but through an absence of wise arbitration.

Part I — The Loop That Digs Its Bed

The Mechanism

The self-amplification loop can be described in a few steps, each of which is a consequence of the previous one.

More data circulate. To process them — for no human window is enough —, one automates. Automation, in order to function, observes: it draws data on its own functioning, on the flows it processes, on the results it produces. And it does not only observe: it multiplies capture points, and it decides — it sorts, routes, prioritizes, triggers —, decisions whose effect marks the real and leaves new traces in it, which new drawings constitute into data. These data join the torrent and call, in their turn, for more automation to be processed. And the loop starts again, one notch higher.

Each turn of the loop digs the torrent’s bed a little deeper. Automation does not merely process the flow: it produces flow, and the flow it produces justifies increased automation, which produces still more flow. This is positive feedback — in the technical sense, in which the effect reinforces its own cause. In the model, nothing stops the loop from within, for nothing in the loop has the function of stopping it; that the real lets it turn in this way is a condition, not a necessity. It is made, one might say, to grow. This loop is the upstream side of a broader closure — the one that passes through reading and through action on the real — whose demonstration belongs to the closing of the tome of the reader, which is its locus; the present tome needs only the return: the effect of decisions comes back upstream, as traces and then as data, and the throughput rises.

This loop extends a movement that From the Notch to the Algorithm had identified: entrusting to the tool what is repetitive so as to free attention for what calls for a decision. But where this movement once freed attention, the contemporary loop saturates it: the automated repetitive itself produces data, which call for new repetitive processing, without the freed attention having time to turn to the decision. The old gesture dug a bed; the loop digs it endlessly.

Why the Loop Is Invisible as a Problem

The formidable character of this loop — the previous document announced it — lies in the fact that each of its steps is, taken in isolation, perfectly reasonable, and that the sum is not. Automating a processing that no human window can take on any longer is rational. Observing the functioning of this automation in order to improve it is rational. Processing the data thus produced is rational. At no moment does one meet an unreasonable step one could point at. The problem is in no step; it is in the sum of the steps, and the sum is no one’s project.

This is what distinguishes self-amplification from a fault or a negligence. There is no one to blame for a particular step, because no step is blameworthy. The difficulty is of another order: locally right decisions compose a globally untenable result. And a result that is no one’s project is a result that no one feels bound to arbitrate.

The loop, however, is not the only engine. It describes how the torrent digs itself deeper, through feedback; it must now be described how it widens, through superposition — and this second engine lies in the way regimes of observation succeed one another.

Part II — The Coexistence of Regimes

A Systemic Leap Is Not Spontaneous

Digging the Canal set out the notion of the systemic leap: the moment when the improvement of the means of a flow does not make the old regime faster or cheaper, but changes the nature of what it makes possible. One would be tempted to picture such a leap as a clean switch: one day the old regime, the next day the new one, the old one abolished. A leap does not necessarily happen like that.

The reason is that the leap does not occur in the abstract, but must be assimilated by situated readers. The Trace and the Reader established it: a situated reader reads only what falls within their window, bounded by what they are. A new regime of observation requires each reader to assimilate the leap to the measure of this window — to learn what the new regime requires, and sometimes to move their window in order to adapt to it. Now this assimilation happens neither everywhere nor at the same rhythm. Generations exposed early to a new regime grasp it easily, as something obvious; those who met it late require, in order to go on acting, that the old services be maintained in their earlier form. Between the two, a transition sets in — and this transition is not an instant; it is a duration, sometimes a whole generation.

Throughout this duration, the two regimes coexist. Society integrates the new flows of observation without abolishing those they were to replace, because part of its situated readers still depends on the old ones. The leap does not abolish: it superimposes. And as long as the superimposition lasts, it is not one but two apparatuses of observation that must be tended for the same object.

Coexistence Adds to Throughput

This coexistence has a direct consequence on the torrent: it adds throughput. And it adds it in three ways, which must be distinguished.

Through the multiplication of formats, first. The same data, of fundamentally identical nature, find themselves multiplied in various formats — one for the old regime, another for the new, sometimes several for the intermediate states of the transition. The same information in substance is inscribed, transported, propagated in several parallel versions, not because it has changed, but because it must remain accessible to readers whose windows differ.

Through the maintenance of parallel paths, next. When several means lead to the same result, the transition does not decide in favour of one alone: it keeps them all, for the time it takes each to find or lose its users. Several routes to a single result are kept active at the same time, and each tends its own flow of observation.

Through keeping things as a fallback, finally. The abolition of a flow of observation is never as clean as one would decide it to be. One keeps the old “just in case”, as a fallback solution should the new one fail. But keeping an apparatus in reserve is not switching it off: a flow of observation maintained as a fallback remains an active flow of observation, which draws, inscribes and occupies its part of the throughput, however little it is called upon.

The most immediate illustration is mail. Postal mail has not disappeared with electronic mail: certain data must still arrive physically, in a form the digital does not replace. Better still, the digital has relaunched it in another format: the contemporary reader receives fewer letters, but many more parcels, whose conveyance is itself steered by a digital flow of observation. The old regime of transport has not been abolished; it has been maintained, moved, and finally fed again by the new one. The underlying point is that the transition between regimes, because it is slow and because it accommodates readers with unequal windows, adds throughputs together instead of substituting them.

A Transition That Never Ends

It remains to measure why this engine, contrary to what its name of “transition” suggests, does not run out. A transition, one would say, has an end: the old regime ends up dying out, and the extra throughput with it. That would be true if systemic leaps were rare and widely spaced. But the torrent is precisely the regime in which leaps follow one another without leaving any transition the time to complete itself. Before the assimilation of the previous regime is complete, a new leap occurs, which opens its own transition and its own coexistence. Superimpositions are not resolved one after another: they pile up. It is no longer a transition, but a permanent regime of transition — and that is why coexistence, far from being a provisional state, is a lasting engine of the widening of the torrent.

Part III — The Diagnosis through the Extended Chain

All the Links Belong to Knowledge

Digging the Canal set out the extended chain, from the flow of the real up to its summit, wisdom: the judicious application of knowledge (Reading, Writing, Understanding). This chain provides the exact instrument for diagnosing both engines at once — for they share, as we shall see, the same defect.

Let us examine them in this light. On the side of the loop: “How can more data be processed?” — a question of knowledge. “How can the functioning of automation be observed?” — a question of knowledge. “How can processing be improved?” — a question of knowledge. On the side of coexistence: “How can the old format be maintained alongside the new one?” — a question of knowledge. “How can the old apparatus be kept available as a fallback?” — a question of knowledge. Each step, in both engines, is a question of how: how to observe more, how to maintain what exists, how to accommodate every window. Knowledge answers these questions admirably, and that is why the loop turns so well and coexistence lasts so easily.

But none of these steps asks the question that only wisdom asks: should we? Should we observe more? Should we automate further? Should we maintain the old regime indefinitely alongside the new, or should we accept switching off a service on which a few readers still depend? Which present interest prevails over the interest to come? Knowledge says how to do what one does; it never says whether one should do it. That question is in no link of the two engines, because they are made entirely of knowledge and wisdom has no assigned place in them.

The Torrent Grows through the Absence of Wise Arbitration

The diagnosis is then clear, and it lies in the very terms that the previous document announced: as long as the two engines described operate without an instance that carries arbitration, the torrent grows for want not of knowledge but of wisdom — through an absence of wise arbitration. It does not swell because knowledge is lacking — it abounds; it is even its abundance that makes the loop turn and that makes coexistence so convenient. It swells because wisdom does not decide, because the question should we? is asked by none of the links that make up these dynamics. Each engine optimizes locally — each step is wise locally — without anything arbitrating globally and over time. To put it in a formula: each step is wise, and the sum is mad.

This result shifts the location of the problem. One might believe that against the torrent one needs more knowledge — better processing, better observations, better means of maintaining the old alongside the new. But more knowledge is precisely what feeds the loop and facilitates coexistence. What is lacking is of another order: an arbitration that would not be one more link in the chain of the how, but an instance of the should we. The torrent is not resolved at the bottom, through more technique; it is resolved — if it is resolved at all — at the top, through the wisdom that decides what is to be done with technique, and what, when the time comes, should cease to be maintained.

Part IV — The Thread of Collective Wisdom

From an Individual Wisdom to a Collective Wisdom

Here opens, in principle, a thread that the last two tomes carry together. Wisdom, as Reading, Writing, Understanding defined it — the judicious application of knowledge —, is understood first in the singular: a person who applies their knowledge, here, now, in a singular situation — a wisdom that is, implicitly, individual. The torrent forces one to think of something else: a collective wisdom. But before naming it, let us measure what the individual can do, and what they cannot — for the second engine, made of adoptions and refusals, gives each person a hold that must be neither overestimated nor denied.

The individual has a hold, and it is elementary. Adopting a new regime or not adopting it is the primary act of which coexistence is made: whoever adopts says yes, whoever does not adopt says no, and the means one stops using loses, in proportion, its reason for being. This hold is real, but its effect counts only when aggregated: an isolated refusal bends nothing; a refusal shared by the great number makes a means useless and forces a change of course. It is a power through federation — not through decision. No isolated observer decides, for the whole torrent, that one must stop observing more or switch off a given old regime; but the great number, through its adoptions and refusals, weighs on it without having decided it.

An asymmetry is then noticeable between the two faces of this hold. In the face of the adoption of a means, abstention does not really exist as a third position: not adopting — whether one refuses, or one is not in a position to adopt, even if one may become so later should the competence become accessible — produces the same effect as refusal. For the means, the silent and the one who declines are equally non-users. Abstention reappears as a distinct position only where the choice is exposed — made visible, stated, weighed in collective deliberation — and where the capacity to adopt is acquired: only then does “I do not adopt, although I could” become a deliberate refusal, and “I abstain from weighing in” a third route, distinct from yes and no.

This asymmetry outlines two levels of decision. At the individual level, the hold is binary and its effect comes from numbers: it is the power that a community federates through its real uses. At the collective level, an instituted system deliberates on the distribution of these choices — how far to constrain adoption, what to protect for those who do not make the transition, which regime to support or to let die out. This level is reflexive: society observes in it what its own choices of observation and adoption produce. It is, in the sense that From the Notch to the Algorithm gave the term, a meta-cognitive act — no longer observing the world, but observing what our observation of the world has made of it — carried this time not by an individual, but by the institutions a society gives itself.

Collective wisdom would therefore be this capacity, no longer of an individual but of a society, to arbitrate present interest against the interest to come. This is the proper contribution of this thread, and it extends the inherited definition without contradicting it: applying knowledge judiciously, but on the scale of a collectivity and over time, where each isolated actor has before their eyes only their own step, locally wise. Present interest acts at both levels: below, whoever has an interest in a regime setting in works to constrain its adoption; above, it weighs on the deliberation that will settle the distribution. Collective wisdom therefore does not arbitrate in a vacuum: it arbitrates the composition of these elementary choices.

Abstention Is Not Neutral

The thread culminates on a point to be stated explicitly, for it is the most important and the least obvious: abstention is not neutral. It is understood here at the level where abstention really exists — the level at which the choice is exposed and the capacity acquired, and not that of private adoption, where not adopting merges with refusing.

One might believe that not expressing one’s wisdom — abstaining, letting things happen, not weighing on collective arbitration — leaves a balanced vacuum, a neutral ground where things would find their own measure. This is false. Not weighing in does not leave a vacuum: it leaves the place to a slanted wisdom. For those who have a present interest will always mobilize; their wisdom will be that of their interest, judicious or not. The abstention of the greatest number therefore does not produce balance: it produces an imbalance in favour of the best-organized present interests.

Collective wisdom is therefore not the spontaneous sum of individual wisdoms, which would come about on its own if everyone kept quiet. It is what comes about — or what is lacking — depending on whether each person gets involved or abstains. In the last resort, it is a matter of involvement. The torrent does not grow only because the loop turns and regimes are superimposed; it also grows because the wisdom that could arbitrate it abstains, and because its abstention leaves the field to the interests that, for their part, never abstain.

What This Thread Prepares

This thread is here a diagnosis: it explains why the torrent grows and where arbitration is lacking. It will be taken up in the next tome as a criterion — for the attitudes the reader can take in the face of the torrent that overwhelms them will be judged, in part, by the yardstick of this collective wisdom. The present tome does not unfold this side: it sets out only that arbitration is lacking, and that its absence is not neutral. The last document of the tome will now lead the dynamic of self-amplification to its end — the threshold where circulation produces more than any window can receive.

Landmarks

For the references used in this document — Descartes’s epigraph on the good use of the mind —, see the Annotated Bibliography of Tome 2. The methodological status is set out in the Preface to the Edition; the loop, the dug bed, the formula “each step is wise, the sum is mad”, postal mail and the parcel belong to the status of illustrations. The concepts taken up refer back to their loci: the torrent, its technical definition and the real-marked-by-us regime, to The Torrent; the movement “entrusting to the tool what is repetitive” and the meta-cognitive act — observing what our observation of the world has made of it —, to From the Notch to the Algorithm; the systemic leap and the extended chain, to Digging the Canal; wisdom as the judicious application of knowledge, to Reading, Writing, Understanding; the situated reader and their window, to The Trace and the Reader.

This document carries, for the trilogy, the locus of self-amplification: the endogenous loop that digs the torrent’s bed — keeping the promise of Digging the Canal, where the loop had only its support: its own amplification is demonstrated here — and the coexistence of regimes that widens its banks, the second engine; the diagnosis through the extended chain — as long as the two engines described operate without an instance that carries arbitration, each step is a question of knowledge, none asks the question of wisdom: the torrent grows through an absence of wise arbitration —, and the thread of collective wisdom: the individual hold that counts only when aggregated, the asymmetry between adoption and abstention, the two levels of decision, abstention that is not neutral. The layout of the looped diagram lives in the Architecture Volume. The closing of the loop — reading that becomes action, the bridge between observation, inference and intervention, the responsibility of the chain — is demonstrated at the closing of the tome of the reader, which is its locus: the present document describes only its upstream side. Everywhere else, these established results circulate without being demonstrated again.

The Self-Knowing Torrent

Speed, Over-Throughput, and the Trace That Threatens at the Edge of the Reader

Tome 2 — Document 8

“Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?”

— T. S. Eliot, The Rock (1934)

Warning

This document closes the second tome, and it occupies a singular place: it is the threshold. The whole tome has built the movement of the datum up to this point; this document leads it to its end, and stops dead where the next tome begins. This boundary is not an effect of composition: it belongs to substance. The tome deals with circulation; it does not deal with reading. So this document will name what threatens at the edge of the reader — but always in the future or the conditional, as a horizon, never as an accomplished fact. What threatens the datum that no one would receive is not ceasing to be what the act made it: it is remaining non-actualized, and losing the value that reading alone maintains; falling back into trace remains the limit case, which this document will never ascertain. Ascertaining it would require a look at reading — and reading is the subject of the next tome. The present document stops at the edge of the reader, and holds this edge rigorously. It is the only point in the tome where a warning note of its own is necessary, because it is the only one where the discipline of the dividing line must be kept sentence after sentence.

Preamble — The Self-Knowing Torrent

The previous document showed how the torrent digs itself, through a loop each step of which is reasonable and whose sum is not, for want of wise arbitration. It remains to lead this dynamic to its end: what, in the end, does a torrent that accelerates without cease produce?

The title of this document indicates the answer, and the shift it makes. The torrent, until now, was described from outside — as a regime one observes. This document describes it from the point where it becomes a problem: that of a finite gaze placed before an infinitely growing throughput. A torrent that knows itself is a torrent grasped again from the situation of the one who must receive something from it, and who discovers that they cannot receive it all. It is here that the tome reaches its limit, and names it.

Part I — The Chain from Speed to the Threat of the Trace

From Speed to Over-Throughput

Let us take up the tome’s established results and tie them into a chain. The faster the datum circulates — through the near-instantaneity of transport (The Properties of Movement). The faster it circulates, the more the real is impacted by it — for the torrent is the real marked by us (The Torrent), and we set its speed. The more the real is impacted, the faster it moves away from the point each act had marked — this is temporal navigation (The Datum Navigates Because It Is Fixed), accelerated by the racing of the real itself. And the faster the real moves away, the more the sources held to be stable lose their stability — for stability (Navigating in Place) was only a ratio of speeds, which the racing degrades.

At the end of this chain, a fact imposes itself, which is the last concept of the tome: over-throughput. Accelerated circulation produces more data in movement than any finite gaze can follow. To this source of throughput — speed — is added a second, which the previous document established: the coexistence of regimes, which multiplies formats and maintains parallel paths for the duration of transitions that never end. Speed raises throughput; coexistence widens it. The two engines of self-amplification converge here, towards one and the same excess. This is not an appraisal; it is a fact of circulation, measurable in principle. There is, at a given instant, a throughput of data that exceeds any capacity of reception. The tome has established the production of this throughput; it is its last established result, and the exact point at which it stops.

Over-Throughput Relative to the Window

This over-throughput calls for careful formulation, for it is the concept at which the tome reaches its limit. The flow of the real has no absolute throughput; it takes on a throughput relative to the fragment observed. Likewise, the torrent does not exceed in the absolute: it exceeds relative to a window. The Trace and the Reader established it: every reader perceives only a band of the spectrum of times, bounded by a threshold of the instantaneous and a threshold of the eternal — they are a situated reader, and their window is finite. The datum that circulates flows in relation to this finite window. When production increases without the window widening in proportion, it is not the window that grows: it is the throughput that rises. And this over-throughput, relative to the window, blurs what one could perceive of the real through it.

Over-throughput is therefore over-throughput relative to the window. It is a fact of circulation, which the tome establishes by argument and not by measurement: in the regime of the torrent, there are more data in movement than any window can absorb. What becomes of what passes before a window without entering it, the tome does not say — for that engages the window, that is, the reader, and the reader is the business of the next tome.

Part II — The Threat Named: The Horizon of Neglect

What the Tome Can Say of the Threat

The Trace and the Reader established the distinction that returns here in full light, and the threshold that holds it: the trace is legible real that is not read; the datum is born of the act of inscription that constitutes it — and between the two, a threshold that nothing crosses on its own, the entry threshold. But there is a second threshold — the exit threshold: from the constituted datum to the datum actualized by a reading — whose edge this tome touches without crossing it; its locus is in the tome of the reader. Circulation leads the datum up to this second threshold; it does not make it cross: a reader is needed.

Over-throughput places the tome just before this threshold — and forbids it to cross. For one can say this, and only this: when circulation produces more than any window can receive — the threshold that The Torrent announced —, part of what circulates risks being actualized by no one. It could pass before gazes without any of them grasping it. What then threatens it is not ceasing to be what the act made it — constituted, it remains so — but remaining non-actualized, and losing, as the real moves away, the value that actualization alone maintains. This is the horizon of neglect: what threatens at the edge of the reader is non-actualization and loss of value. Falling back into trace remains the limit case — the limit-ideal in which even the conditions of any interpretation would be lost — which this tome does not have to measure: the tome of the reader, from its opening, fixes its scale and its measure. Over-throughput produces, potentially, the conditions of neglect.

But the tome cannot say of a datum that it is neglected — still less that it has fallen back into trace. Saying so would presuppose ascertaining that no reader has actualized it or will actualize it — a finding that concerns reading, and therefore exceeds circulation. The tome holds to the future and the conditional: the datum could never be actualized; it threatens to lose its value; circulation produces the conditions of neglect. It never ascertains abandonment. This tipping over — from threat to fact — belongs to the reader, and the reader belongs to the next tome.

Why This Restraint Is Not a Stylistic Caution

This restraint is not a scruple of formulation; it is commanded by the very thesis of the tome. The thesis of the tome is: the datum navigates because it is fixed. The tome has gone as far as the point where navigation becomes so fast that it exceeds any window — and it stops there, because beyond, it is no longer navigation that is at stake; it is reception. Crossing the threshold would mean abandoning the thesis of circulation for a thesis of reading, which is not its own. The tome keeps its thesis intact by stopping exactly where circulation gives way to reading.

This is what makes this point a threshold and not a mere end. An end closes; a threshold opens onto something else. Over-throughput is the last word of circulation and the first problem of reception. The tome sets it out as the last word; it leaves it, as the first problem, to whoever will take over on the other side.

Part III — What the Threshold Leaves to the Next Tome

The Question Turned Around

The tome has described over-throughput from the side of production: circulation produces it; it is an objective fact. There remains another face, which the tome glimpses without dealing with it: over-throughput from the side of reception — a finite gaze that undergoes it. These two faces are those of a single concept, and this is what makes it a weld between the two tomes rather than a border. The datum which, here, is potentially neglected because the throughput exceeds any window will be, on the other side, either actualized by a situated reader, or neglected in fact — non-actualized, of a value that nothing maintains any longer, falling back into trace remaining, at the very end, the limit case. The tome never crosses this either… or; the next tome will deal with nothing else.

The question the tome leaves open is therefore formulated thus: what becomes of what arrives without being read? And: what can a reader who knows they are overwhelmed do? These are not questions of circulation — circulation has done its share; it has carried more than any gaze can follow. They are questions of reading. The tome asks them as one holds out a hand across a threshold, without crossing it.

The Safeguard Handed On

One thing, however, must cross this threshold without being diluted, for it holds on both sides: the torrent is our resultant, not a natural fatality — The Torrent established it, and drew the consequence from it: because it is ours, there is something to answer for. Let us say one thing about it again at the moment of handing over, because what follows depends on it: over-throughput inherits this signature. The reader who knows they are overwhelmed is not a victim of the rain; they are a portion of what dug the torrent. This changes nothing in circulation, which has done its share. It changes everything in what the next tome will be able to require of the reader — but it is for the next tome to say so.

The Last Sentence of the Tome

The Datum Navigates Because It Is Fixed opened this tome on an idea: the datum is the means humans have given themselves for not letting everything be carried away by the flow of the real — the hold they tear from the current so that a fragment of the present may abide and be transmitted. The tome ends on the turning around of this means. For what was to hold the present against oblivion now produces more data than any gaze can follow; and what arrives without being read remains mute — held, fixed, constituted, but of a value that reading alone would maintain. The datum navigated because it was fixed; it navigated until it exceeded every window; and at the end of its navigation, it waits, within reach of gazes that cannot all follow it, for a situated reader to actualize it — or for none to do so, and for neglect to draw it towards the horizon where it would carry nothing but its inscription. The instrument humans had given themselves to hold the world threatens, through sheer abundance, to let it slip away again. What will be decided there is not decided in circulation. It is decided in reading. And reading begins where this tome ends.

Landmarks

For the references used in this document — Eliot’s epigraph on wisdom lost in knowledge —, see the Annotated Bibliography of Tome 2. The methodological status is set out in the Preface to the Edition; the self-knowing torrent, the edge of the reader and the hand held out across the threshold belong to the status of illustrations. The concepts taken up refer back to their loci: the distinction between the trace and the datum, the entry threshold that nothing crosses on its own, the situated reader and their window as a band of the spectrum of times, to The Trace and the Reader; the near-instantaneity of transport, to The Properties of Movement; temporal navigation, to The Datum Navigates Because It Is Fixed; stability as a ratio of speeds, to Navigating in Place; the technical definition of the torrent, the real-marked-by-us regime and the responsibility that follows from it, to The Torrent; the two engines of self-amplification, to Self-Amplification. The formalization of over-throughput — arrival rate, capacity, queue, deadline — lives in the Architecture Volume.

This document carries, for the trilogy, over-throughput relative to the window — the last established result of circulation: in the regime of the torrent, there are more data in movement than any window can receive — and the locus of the threat of the trace, scoped as the horizon of neglect: what threatens at the edge of the reader is non-actualization and loss of value; falling back into trace remains the limit case, whose scale and measure belong to the tome of the reader, which opens on them. The threshold between the two tomes belongs to substance — the line where circulation gives way to reading — and it commands the discipline kept here: the document never ascertains abandonment; it names it as a horizon, in the future and the conditional. The exit threshold — from the constituted to the actualized — and what a reader who knows they are overwhelmed can do are the business of the tome of the reader. Everywhere else, these established results circulate without being demonstrated again.

Thus ends the second tome, From Spring to Torrent — The Movement of the Datum Tried by the Flow. The movement of the datum has been led in it from the act of exchange to the over-throughput that exceeds every window. What becomes, beyond this threshold, of the datum that arrives without being read — actualized by a situated reader, or neglected up to the horizon where the trace would regain its rights — is the subject of the third tome, The Trace or the Datum — The Situated Reader Tried by the Flow.

Annotated Bibliography of Tome 2

Apparatus of References of Tome 2

This bibliography brings together the references proper to the second tome of the essay. It extends the Annotated Bibliography of Tome 1, to which it refers for the whole common foundation (Bergson, Whitehead, Peirce, processual ontology, the history of the datum, literacy). It does not repeat its content: it adds the references that the movement of the datum — the act of exchange, transport, propagation, the torrent — calls for in its own right. Like that of the first tome, it is organized to remain useful over time.

What Will Not Change

At the Philosophical Foundations

Whitehead, Alfred North — Process and Reality, 1929. Already foundational in the first tome, he is doubly so here: the definition of the datum as “what is given for a later occasion” is the direct underpinning of the thesis that circulation is a frequent and structuring regime of the datum, and that transport extends its availability and its possibility of use. It is from this formula that the first document of the tome starts.

Mauss, Marcel — Essai sur le don. Forme et raison de l’échange dans les sociétés archaïques, 1925 (first published in L’Année sociologique). The triad “to give, to receive, to return” is called upon as the intuition of a social gesture that calls for a recipient and engages reciprocity. The tome retains from it the idea that exchange is never a mere neutral transfer, without calling on its complete anthropological apparatus. A methodological caveat: the datum is not ceded the way the Maussian gift circulates from hand to hand, since it is duplicated; the analogy illuminates the act of exchange; it does not describe it exactly.

Bergson, Henri — L’Évolution créatrice, 1907. For duration as continuous flowing, the underpinning of “temporal navigation”: the real moving away from the marked point while turning around it. See the Annotated Bibliography of the first tome for the detail.

Political Economy of Circulation

Ostrom, Elinor — Governing the Commons, 1990. The theory of the commons, already called upon in the first tome for the flow, sheds light here on the political dimension of transport: irrigation has its beneficiaries and those it leaves out, and the circulation of the datum raises questions of collective governance that cannot be reduced to the private ownership of an object.

Zuboff, Shoshana — The Age of Surveillance Capitalism, 2019. For the contemporary issues of massive capture and circulation. To be read, in accordance with the Preface to the Edition, as a dated illustration of the forms the torrent takes, not as a lasting description.

Theory of Transmission and of the Signal

Note on the materiality of digital transport — The assertion that every digital datum in movement is, at every instant, carried by a physical medium (fibre, wave, buffer memory) is a working hypothesis of the essay on the physical condition of networks, which it does not demonstrate; the body of the text states its consequences under this condition. The essay retains from it the principle of substance — there is no immaterial transport — without calling on its formalism. For the environmental footprint of transport and storage infrastructures: International Energy Agency (IEA), reports on the consumption of data centres and networks.

Note on the “systemic leap” of instantaneity — The disappearance of the proper duration of transport (the passage of delay below the threshold of perception) is analysed in Digging the Canal, under the systemic leap, and taken up here. For the professional notion of perishable data and that of data freshness, see the corresponding note in the Annotated Bibliography of Tome 1.

Sources of the Epigraphs

Whitehead — Process and Reality, 1929, for the epigraph of The Datum Navigates Because It Is Fixed (entry above).

Ovid — Epistulae ex Ponto, I, 5, lines 5-6: “ut capiant uitium, ni moueantur, aquae” (waters go bad if they are not set in motion); the epigraph of The Datum Navigates Because It Is Fixed is an adaptation of it.

Mauss — Essai sur le don, 1925, for the epigraph of The Act of Exchange (entry above).

Socrates — the epigraph of Transport and Propagation, given under attribution: the formula circulates under his name with no direct source, Socrates having written nothing. The attribution is kept for what it is — a usage, not a reference.

Jefferson, Thomas — letter to Isaac McPherson, 13 August 1813, for the epigraph of The Properties of Movement. The letter is the classic source of the idea that a transmitted idea does not leave the one who gives it; the tome makes it the precedent of copying without division.

Naisbitt, John — Megatrends, Warner Books, 1982, for the epigraph of The Torrent — the abundance of information and the hunger for knowledge. Single locus of the duplicate: the tome of the reader refers to it without repeating it.

Descartes, René — Discours de la méthode, 1637, for the epigraph of Self-Amplification.

Eliot, T. S. — The Rock, Faber & Faber, 1934, for the epigraph of The Self-Knowing Torrent.

Textual Scholarship and Diplomatics

Two traditions of form, called upon without being developed: this tome is the locus of filiation and provenance, and their vocabulary comes from there.

Mabillon, Jean — De re diplomatica libri VI, 1681. A treatise on the critique of the authenticity of documents: the historical anchoring of what the tome calls constitutive provenance.

Textual criticism — Lachmann, Karl, for the genealogical method; Bédier, Joseph, La Tradition manuscrite du « Lai de l’Ombre », Romania, 1928, for its contestation. The vocabulary of editing — filiation, critical apparatus, witness, contamination — is borrowed from this tradition, not its doctrine: the tome retains that the descent of copies is documented and discussed.

Neighbours of the Thesis

The Architecture Volume carries the map of neighbours (§7); one family touches on the proper subject of this tome: its references are below; those of the other families are in the Annotated Bibliography of Tome 1 and the Annotated Bibliography of Tome 3.

Trajectories and materiality — data journeys — Bates, Jo, Lin, Yu-Wei & Goodale, Paula, Data journeys: Capturing the socio-material constitution of data objects and flows, Big Data & Society, vol. 3, no. 2, 2016; Leonelli, Sabina & Tempini, Niccolò (eds), Data Journeys in the Sciences, Springer, 2020 (open access). Overlap: material movement, transformation, reuse, change of context, and the friction that slows them. Fruitful tension: circulation does not always preserve an invariant datum — it can manufacture the object that circulates. This is what thesis 9 of the Architecture Volume integrates by distinguishing preserving, altering, engendering.

Internal Cross-References of the Essay

The second tome relies continually on the first. The following concepts are established there and are not redefined here:

• object ontology and its critique, the datum as a constituted entity of which drawing is the observational route, the inappropriability of the flow and the ownership of the drawing, the backup metaphor, the three ages of the datum — The Flow and the Annotation;
• the flow of observation, the observational optimum, the extended chain (flow of the real → flow of observation → elementary data → consolidated data → information → knowledge → wisdom), the systemic leap, the LLM as the fusion of the library, the librarian and the storyteller — Digging the Canal;
• the two times of the datum, the act of inscription and fixation, the two deaths (lapse and disappearance), the inaccessible datum, the datum carried and not ceded, freshness that cannot be recovered, the image of the mobile frame that moves away around a fixed object — The Two Times of the Datum;
• the legibility of the real, the caused trace opposed to the intended datum, the situated reader, the window as a band of the spectrum of times, humans as a portion of the flow, the wave and the ocean — The Trace and the Reader;
• the torrent, the rule that points out the shoals without describing them — The Wave and the Ocean;
• the “privilege → common” movement and the history of libraries, the movement “entrusting to the tool what is repetitive” — From the Notch to the Algorithm;
• the DIKW pyramid and wisdom as “judicious” application, managerial co-optation — Reading, Writing, Understanding the World through Data;
• the substance/form distinction, the status of illustrations and guiding images, the status of the “flow of the real”, the role of AI — Preface to the Edition.

Concepts Introduced by the Present Tome

The reference definitions are in the Canonical Lexicon for the concepts that have an entry there or a correspondence of heading; the groups, images, sub-concepts and articulations named here for the record have no entry of their own there and are demonstrated in the document indicated after each element, the precisions they carry belonging to the same document. For the record, the concepts proper to the second tome are: the act of exchange (symmetrical to the act of inscription), in The Act of Exchange; the transport medium and its minimal typology (moving, transmitting, making accessible), in The Act of Exchange; double navigation (spatial and temporal), in The Datum Navigates Because It Is Fixed; the image of the spiral, in The Datum Navigates Because It Is Fixed; transport as a relation between two movements and its criterion of accomplishment (the reader’s share made practicable), in Transport and Propagation; the inaccessible datum as an unaccomplished transport, in Transport and Propagation; propagation as the aggregate of transports, in Transport and Propagation; the properties of movement (fidelity, temporality, multiplication, and the three axes of duplication), in The Properties of Movement; transport probity, in The Properties of Movement; the abolition of transport ageing by instantaneity, in The Datum Navigates Because It Is Fixed; stability as a ratio of speeds (navigating in place), in Navigating in Place; the torrent as the regime of the real marked by us, in The Torrent; self-amplification and its diagnosis through the extended chain, in Self-Amplification; collective wisdom and non-neutral abstention, in Self-Amplification; the aggregated individual hold (power through federation), in Self-Amplification; the two levels of decision (binary individual and reflexive collective), in Self-Amplification; over-throughput relative to the window as the threshold of the tome, in The Self-Knowing Torrent; irrigation, in Transport and Propagation.

The work of opening paths, in Transport and Propagation — the third document —, where it is sketched as a consequence of the impediment of transport, voluntary (the secret) and involuntary (incapacity), receives its scope there: remedying incapacity, without forcing secrets. It will be the subject of a separate elaboration, distinct from the essay, around a data craft.

Translation choices

These notes say how to read the English words that render the terms of the corpus; they are fixed for the whole English translation by the Translation Charter of the Corpus. The notes of the first tome remain valid here; those below repeat the main ones and add the terms of movement.

• datum / data. The corpus treats the datum as an individual: constituted by an act, dated, carried by media. The translation therefore says datum in the singular — the Latin for “what is given” — and data as its plural, construed as a plural. Whitehead’s datum, in italics, is the philosopher’s term, distinct from the datum of the corpus.
• drawing / to draw (prélèvement, prélever): the gesture of drawing from the flow, as one draws water; sampling is avoided. To measure, entry and survey render relever and relevé; reading is kept for the act of reading.
• lapse; stale; to go stale (péremption; périmée; se périmer): a relation, not a decay — the gap that grows between a datum that stays still and a world that carries on.
• spring / stable source (source). French has one word where English has two. In the image of water — the title of this tome, the springs that feed the canals — source is a spring; the stable sources are the fragments of the real that lapse so slowly that one can build on them. Source keeps its ordinary sense elsewhere.
• transport (transport, transporter) is never the shipping of goods, which the text calls physical transportation or mobility. Transport is not the movement of the datum but a relation between two movements — the datum towards the reader, the reader towards the datum. Movement renders both mouvement and déplacement. The transport medium (support de transport) is what moves, transmits or makes accessible; the medium (support) carries the inscription; a carrier (porteur) holds an occurrence of the datum.
• transport ageing / ageing of origin; transport freshness / freshness of origin (vieillissement de transport or de trajet; fraîcheur de transport): the delay of the journey, against the age the datum has from its act.
• the real-without-us / the real-marked-by-us (réel-sans-nous, réel-marqué-par-nous): the two regimes of the flow; the torrent is the second.
• throughput; over-throughput (débit; sur-débit): the quantity of data in movement per unit of time, and the excess of that quantity over what a window can receive. Never overload or overflow: over-throughput is a fact of circulation, relative to a window.
• neglect; neglected datum (délaissement; donnée délaissée): the datum that no reader actualizes. Falling back into trace (retombée en trace) is its limit case, which the tome names without ascertaining it.
• actualized (actualisée): the act by which a reading produces a meaning, never the updating of a file. The entry threshold leads from the trace to the constituted datum; the exit threshold is what follows reading without being part of it: inscribing, calculating, saying, acting — or nothing. The Self-Knowing Torrent also uses it for the passage from the constituted datum to the actualized one; the translation keeps the term.
• canal; to canalize (canal; canaliser): the canal dug in the river, for irrigation; never a channel.
• recipient, receiver, sender render destinataire, récepteur, expéditeur.
• opening of paths; impediment (ouverture des chemins; empêchement): the work that remedies incapacity without forcing secrets.
• individual hold; power through federation (prise individuelle; puissance par fédération): the individual’s decision, which counts when aggregated.
• probity (probité) is the honesty proper to each act: of drawing, of transport, of the reader.
• Tome names each of the three tomes; volume, any other book of the corpus. “The tome of movement” and “the tome of the reader” are Tomes 2 and 3. Threshold (Seuil) is the opening text of a tome, and the passage an act makes one cross; Landmarks (Repères) is the closing section of each document; locus (plural loci) renders foyer, the single place where a contribution is demonstrated.
• Quotations. Jefferson, Naisbitt and Eliot are given in their original English; Descartes in the received English translation of the Discourse on the Method. The epigraph attributed to Whitehead condenses his use of the word datum; no sentence of Process and Reality matches it word for word, and it is translated from the French. The epigraph after Ovid is translated from the French adaptation, with the Latin line beside it. Mauss’s triad and the sentence attributed to Socrates are translated from the French.
• Titles. Tried by the Flow renders À l’épreuve du flux: the datum put to the test of the flow. This tome is From Spring to Torrent — The Movement of the Datum Tried by the Flow.