Power
In Alexandria, power is not something one party does to another. It is a condition between two structures, present whether or not anyone acts, decides or notices. (Treatise 10.2, 10.8)
Two structures, and the relation already holds.
Until now there was one structure, yours or civilisation’s. Now there are two: call them A and B. They could be you and another person, you and a company, two organisations, two groups or two countries. A, here the larger of the two, and B stand for all of these.
Neither knows the other exists, yet what A has already worked out shapes what B can reach, wherever A got there first; and the same holds the other way wherever B did. Nothing has happened, and nothing has been added to the theory to make it so. (Treatise 1.2, 10.5, 10.30)
Alexandria defines power in two parts: what each holds, and what one determines for the other.
A’s power to is the structure A holds, with the vantage that structure gives it, and the same is true of B. Power over is the relation in which one structure determines, to the extent that it does, what the other can search, read, commit toward or cross. It acts on what is available to B, never on what B does. It needs no contact between them: it is a condition, not a connection and not an act. (Treatise 10.8, 10.9, 10.30)
B notices A.
A’s settled structure gives off a signal: physical presence, market presence, a track record. B orients toward it, and that orientation launches B’s search. This route opens rather than closes. A gives off light from what it is and what it has settled, to anyone who can observe it from where they stand. It is not aimed; it simply emits. (Treatise 10.4, 10.7, 2.77, 2.118)
B reads what A has made visible.
Its structure, its public accounts, what it has published. What A has submitted is open ground, and B reads it on arrival: this is retrieval, not discovery. What B sees is never the whole of A. What lies beyond B’s vantage is not hidden. It is simply not visible from where B currently stands. (Treatise 2.17, 2.119, 2.125)
B searches further into A’s structure, and some searches come back empty.
B now meets two limits, and it experiences them very differently. The first is what A holds and keeps to itself: trade secrets, internal data, private terms and A’s own intentions. These sit deep within A’s structure, or behind a barrier A has put in the way, and they do not return, in whole or in part. B may know that something is being withheld, or may not. (Treatise 10.9, 10.10)
The second limit: the return arrives, and B cannot make sense of it.
Patents, arrangements, technical filings and contracts come back from A’s structure and arrive whole. B can read them, but lacks the context, expertise or understanding to make sense of what they fully mean. This is a by-product of A’s confirmed structure, and nobody has to cause it; the limit is B’s own structure, and what B can make legible at that time. Essentially, B does not know what it does not know. (Treatise 10.10, 2.68, 3.50)
B finds the position it wants, and sees that A stands on it.
It might be a market segment, a site, a channel or a licence. B does not have to go there to be constrained by it: knowing the ground is taken already shapes where B will commit. That knowledge was taken on, not tested, so the ground stays open to challenge, and only B going there would show what is really held, and to what extent. (Treatise 10.4, 10.6)
Held, provided or submitted: how B could be there, and B usually learns which only by asking.
If the position is held, B can still go there, but where A really stands, B finds nothing A did not reach first. To take that ground, B would have to show that A’s position cannot hold, and A will usually see that move. If it is provided, B can be there with A’s concurrence, on terms A can set and withdraw. If it is submitted, it is highway anyone may use, and no one, A included, can close it. B finds out which when terms come back, or when they do not. (Treatise 10.6, 10.10, 10.14, 10.35, 10.40)
B weighs how strongly A would defend the position, and it looks costly.
Whether or not it is. A deeply settled position gives off light from every surface, and B prices the approach as expensive: A’s record of counter-moves, a reputation for litigation, or sheer scale. Nothing false need be shown; what is there is displayed so that it will be read.
This is the cheapest route of all because the cost exists only in B’s reading. B takes on, without testing it, that the direction is too expensive, and commits elsewhere. A needs nothing standing at the position to make that work: no protection, not even occupancy. It holds until someone goes anyway, and the first party to pay finds out what the price really was. (Treatise 10.6, 10.9, 10.10, 2.77)
Getting to the position actually costs more.
A has built protection around it: supporting structure, a patent thicket, an exclusivity agreement, a price.
This route works only if B reads it. A price deters only a party that knows the price, and a patent thicket deters only one that has found the patents. Every other closing route works with nothing for B to read: what is obscured never arrives, what is illegible cannot be read, and a displayed cost is read as the true price of the ground. An imposed cost is the one place where B must see that A is constraining it, and a party that sees the constraint can push against it. That is why it is the route that provokes resistance. (Treatise 10.10, 10.13, 2.134)
B cannot tell the two apart from where it stands.
From B’s vantage, how defended the position looks and how defended it is read as one. Only a crossing (a traversal) exposes the real price, and a displayed cost collapses the moment anyone pays it. What exposes it is B’s own commitment, not A’s reaction, which would belong to strategy. There are two cheaper tests short of committing: the record of what others who tried were charged, and pricing the direction explicitly beforehand. (Treatise 10.14, 10.103)
B chooses.
B can coexist: alongside A’s structure, on ground A neither holds nor is reaching for, or on A’s own terms. The relation stays a standing condition, and this is still power. If A then moves against B there, it is A that has begun a contest, which is the subject of Part 2, Strategy.
Or B can compete: commit toward ground A is reaching for, or show that A’s position cannot hold. Competing is what B chooses; contest is what it becomes once A can read the move, whether or not A has responded yet, and the sequence continues in Strategy. If A cannot read the move, it remains a relation of power until A can. (Treatise 10.14, 10.17, 10.98)
If B coexists on A’s terms, a dependency builds without anyone acting.
B builds on the access, and what it builds bonds to what A provides. The more B builds, and the better it uses the access, the more a withdrawal would cost it, so A’s leverage is made by B’s own success. Each concession is small, until complying becomes the rational choice, and A rarely needs to withdraw at all.
B can limit the exposure by keeping what it builds on the access shallow and separable, at the price of the efficiency that deeper use would bring. The only way out is to build for itself what the provided position supplies. A, losing its leverage, may respond, and that response would be strategy. (Treatise 10.40, 10.42 to 10.44)
Power is read at an instant, and it changes with every arrival.
Every route in this Part acted on something B lacked: ground not yet reached, structure not yet held. Fill the gap and the route has nothing to act on. But the gap never stays still. Each time anyone confirms something, A, B or a third party, the frontier moves and what each holds changes, so the relation between A and B grows or shrinks with that event. It has no stable level, and what this Part has shown is the relation at one moment.
And because each discovery extends the frontier rather than reducing it, power never goes away. Only submission removes a position for good from what can be closed against anyone, and it does so one position at a time. (Treatise 10.31, 2.17, 4.39)
The whole of it, seen along B’s line of sight.
Flattened along B’s line of sight, everything in this Part fits on one page: A above, B below, and the six routes where B meets them. Four close what B can reach and two open it: luminosity, and occupation where a position is submitted as highway. There is no contact between A and B. Obscuration and illegibility sit in A’s structure, where B searches and reads; both forms of suppression sit at the position B wants; occupation is A standing on it. (Treatise 10.10, 10.12)
Strategy
Power holds whether or not anyone acts. Strategy begins when someone does: it needs an actor, an intent, an act and a target, and it comes in episodes where power is standing. Power does not stop while it runs. At every moment of a contest the relation stands as the acts so far have left it, and a contest halted at any point leaves that state standing as power. (Treatise 10.2, 10.98)
Ground both are reaching for.
Between A and B lies ground neither has yet reached and both have claimed: absences at the edge of each structure, in the same part of the unknown, so that whichever confirms there first moves the boundary of both. Perhaps a market both are entering, a technology both are developing, or a region both want to influence. Nothing has happened yet; this is still power. (Treatise 10.98, 7.8)
It begins with a move the other can read.
B commits toward that ground. If A can read the move, contest has begun, whether or not A has responded. If A cannot, it remains a relation of power until A can. From here, what each side does is aimed: it has an actor, an intent and a target, and it acts on the other’s traversal of the contested ground. B has made the first move, and what follows is A’s response: the established structure acting on the newcomer. The direction is the same as in Part 1, A shaping what is available to B, now done deliberately. Step 33 turns it around. (Treatise 10.98, 10.17)
What no act can reach.
A cannot reach B’s search: it is launched from where B stands, bounded by what B holds, and nobody else can launch it. Nor can A reach what B is able to read, because that depends on B’s own structure. What A can reach is everything that arrives at B from outside, and the ground between them. Every act in a contest lands on one or the other. (Treatise 10.98, 2.67, 3.50)
The two things an act can land on.
B’s positions are what B has built, standing in the territory: they are built, occupied, dislodged and defended. B’s reading is what B takes the terrain ahead to be: where A stands, what is open, what a direction will cost. In the figure it is drawn as a small map beside B’s vantage. It is contested by arranging what arrives.
The two are joined: B commits from its reading, and commitment builds B’s positions. So an act on one relocates the other, and contest is a process rather than an exchange, dated at every step. (Treatise 10.99, 4.39)
B’s centre of gravity need not be its strongest point.
Every structure has a position on which more of the rest depends than on any other. Sever it and what had resolved against it reopens, layer by layer outward. This is what Clausewitz called the centre of gravity: the hub on which everything depends, not the place where strength is concentrated. B reinforces what it values or fears losing, and that is independent of what it depends on: the two can coincide, and often do not.
An act against B’s most reinforced position pays the highest price and runs a local cascade; an act against its centre of gravity, where the two differ, costs less and reopens far more. Getting this wrong compounds, because what was spent on the strong point is not available where it would have counted. And the analysis that finds the other’s load-bearing position finds your own. (Treatise 10.100, 8.225, 2.108, 2.109)
Beneath the threshold of major response.
A can also take ground in small, confirmed steps, each calibrated so that the cascade it runs in B stops short of the point at which B would have to treat A’s actions as demanding a reply. The threshold is the size of the cascade a move would run in B, and nothing else: not visibility, not category, not any line drawn by convention. Each step moves the boundary for the next, so this kind of contest is patient by construction, and a reply in kind is itself an incursion that invites the response it was meant to end. (Treatise 10.100, 8.248)
Five routes, one object.
B’s reading is formed from what arrives at B’s vantage, and A can act on what arrives in five ways: obscuration, deception, staleness, suppression and saturation. None of them touches the instrument that forms the reading. All five land on the same object, which is why they belong together. (Treatise 10.101, 8.226)
A withholds, and B’s search comes back empty.
Nothing false is introduced. What B would have found never arrives, and B’s vantage stays where it was, because the structure that would have moved it never came. In Part 1 the same limit could obtain with nobody doing anything; here A deliberately obscures part of its confirmed structure from B. (Treatise 10.101, 10.3, 2.125)
A sends B something false that appears true.
Something that is not there returns to B as if it were. A does not need to change its structure to do this; it changes what returns. B’s reading now shows a position where there is nothing, drawn as a dashed outline, and B’s next commitment is formed against a surface that does not exist, which is what A intends. Deception is not displayed cost. Display presents what is there so that it will be read; deception presents what is not, and fails the moment anyone tests it. (Treatise 10.101, 4.41)
A changes the ground faster than B can re-read it.
A builds into ground that B’s reading still shows as open. Nothing is concealed and nothing is false: B’s reading was sound when it was formed, and it is now out of date. It is the hardest route to detect from inside, because every individual reading was correct when it was made. (Treatise 10.101, 4.39, 2.102)
A raises what B’s direction costs, or what it is read to cost.
Imposed: A builds protection on the approach, so the direction actually takes more, and B can assess that it does, until B narrows its own commitment for sound reasons. Displayed: A’s resolved position emits from every surface, and B prices the approach out of its planning whether or not it would cost that much. These are the routes of Part 1, now done deliberately and aimed at B. (Treatise 10.101, 2.134, 2.77)
More arrives than B can read.
A cannot make B unable to read; that bound is B’s own structure. But B’s capacity to make a reading from what arrives has a limit, and A can exceed it with material that is not concealed and is designed to overwhelm, arriving faster than it can be made into anything. The material need not be false; what does the work is the volume. Where some of it is false, deception is running alongside. Sound returns lie unread. (Treatise 10.101, 2.67, 3.50)
Where an act becomes a condition.
Two of the five, obscuration and suppression, are also routes of power, because each can be done or can simply obtain. Three, deception, staleness and saturation, are acts only: each needs someone doing it.
Staleness shows where strategy succeeds. Once what A built into the contested ground is confirmed and held, nothing more needs doing: the act has become occupation, a condition of power. A successful strategy is one that no longer needs doing. Strategy is one of the ways power comes to obtain; it is not what power is. (Treatise 10.3, 10.101)
Acting through others: vantage does not transmit.
When A acts through others, its people, its units, its partners, a reading formed at the centre cannot be installed in a member who did not form it. Every instruction is also dated: the route it specifies runs through terrain the other side is changing, and referring back for fresh instructions produces staleness of its own.
What survives transmission is the absence being approached, not the path. So the centre states the absence, and the member asks what would advance it from where they stand. That works only if the member holds enough structure for its reading to be worth trusting, if the centre can state an absence rather than a route, and if both keep a record that can be checked. (Treatise 10.102, 4.39, 7.8)
Defence is one activity.
Because all five routes land on one object, B’s defence is one activity in different proportions, not five. Extending the search answers obscuration. Auditing what arrives answers deception. Re-reading at a tempo matched to how fast the terrain is being changed answers staleness. Pricing a direction explicitly separates imposed cost from displayed. Deciding in advance what will not be read is the only answer to saturation, and the one defence that requires accepting a loss. (Treatise 10.103, 2.164, 10.118)
It detects contest before its route is known.
A reading predicts what a direction will demand; crossing it returns what it actually demanded. Keeping both side by side, B can see that it is being acted on before it can say how, because the gap between predicted and returned shows in the record whether or not the route has been identified.
The same record locates a failure afterwards: in the reading, in the choice of end, or in the route between them, three different faults with three different remedies. A structure that keeps no record experiences contest as bad luck. (Treatise 10.103, 2.70)
A crossing toward a distant absence, while another structure works to prevent the confirmation that would reach it.
That is Alexandria’s definition, and it gives the familiar frame its object. Ends are the absences being approached. Ways are the crossings, and the routes acting on them. Means are the structure each side holds. Ends, ways and means described a traverse with the traverser left out.
Business strategy, military strategy and assessment grew up as separate literatures because each stood nearest a different object: positions, readings, or both. What is established in one carries to the others without translation. (Treatise 12.27, 8.200, 10.104)
Contest runs both ways.
Everything A has done to B in this Part, B can do to A at the same time: act on A’s reading, look for A’s load-bearing position, take ground beneath A’s threshold. And a newcomer that commits where others declined is what undoes a displayed cost. Each side is acting and acted on at once, and each act changes the terrain the other’s next reading is of. The analysis that finds the other’s load-bearing position finds your own, by the same reading turned inward. (Treatise 10.98, 10.99, 10.100, 10.6)
The whole of it, seen along B’s line of sight.
Flattened as in Part 1, now with A acting on B: the contested ground between them, the five routes landing on B’s reading, the acts on B’s positions and, below, B’s record of what each direction was predicted to demand against what it actually demanded. (Treatise 10.98 to 10.103, 2.70)
The personal register
Every relation in the first two Parts reaches a person, or it reaches no one. A gate, a disposition or a record acts only by determining what some particular person can search, read, commit toward or cross. This Part follows the relations to that point, and finds one route the earlier Parts could not express. (Treatise 10.70, 3.73, 7.25)
Now B is one person.
A is everything around them: the institutions they pass through, the organisations they work in, the platforms they use, the society they live in. The structures are the same as before, and so are the routes. What changes is that B is a single traverser, and everything that follows arrives at one place. (Treatise 10.70)
Nobody can move their own vantage, but another structure can shape where it ends up.
Where B stands follows from what B holds, so B can only build, and stand where their confirmed structure puts them. A cannot move B’s vantage either, but it can determine what B builds: what is there to be retrieved, and what B is required or enabled to cross. A role, a school, a stream, a posting, an apprenticeship, a neighbourhood, a language of instruction.
No single placement closes anything. One school, one posting, one first job: none of them shuts a door. Together they decide what a person has in front of them for years, and so what they can make sense of afterwards. That is why any account that looks for an act never finds this route. It is also the only route that works on what B holds, B’s power to. Every other route works on what B can do from where B already stands. (Treatise 10.71, 10.72, 2.125)
Placement is read as merit.
Its effects are slow, cumulative and lasting, the way structure is: a vantage arrived at over fifteen years is not undone by learning that it was assigned. And they are indistinguishable, from outside and from inside, from the person’s own capacities. Someone who cannot read what a colleague reads easily shows the same limit whether it came from where they were placed or from anything else. That is why structural advantage is so consistently read as merit, by those who have it and by those who do not. (Treatise 10.76, 3.50)
Access is easily mistaken for possession.
Much of what a person relies on is provided: an employer’s systems, a platform, a licence, a professional body’s standing. In use it is indistinguishable from what is theirs. The test is simple and rarely applied: could B rebuild it if the provider vanished tomorrow? If so, what B holds is highway, and the gain is real and permanent. If not, what B holds is access, and the dependency described in Part 1 is building with every competent use. (Treatise 10.73, 10.74, 10.40)
Attention follows the light, and the light follows what is already there.
People work where they can see something to work on. A field with a large body of confirmed work is easy to see: it has an established literature, known problems and a next question in plain view. A field with little confirmed work offers none of that. So people commit toward the well-lit fields, those fields grow denser, and research, professional attention and investment follow a gradient nobody set. In Alexandria’s terms, only settled structure emits, so light concentrates where the mass already is. This is the field effect Bourdieu described, and Alexandria supplies its mechanism. Nothing pulls on anyone; what bends their path is what is visible from where each has been placed.
A dark region is not less likely to hold something worth finding. It is less likely to have been crossed, which is a fact about history, not about the ground. And the better the system works, the more strongly it passes those regions by, because every confirmation adds weight where weight already was. (Treatise 10.81, 10.82, 10.83)
Reading a certifier’s standing as a warrant about the thing certified.
Almost everything anyone holds was accepted from someone else, and that is unavoidable. Here B accepts a position because a structure with real standing has certified it: an accrediting body, a regulator, a journal, a ratings agency, a firm with a long record. Its light is genuine. What it tested was compliance with a stated standard; what B reads is that the thing is sound.
The gap between the two is not hidden. It is written down in the standard, and almost nobody reads it, because the point of relying on the certificate was to avoid that work. No deception is needed, and usually none occurs. (Treatise 10.84, 10.85, 10.86, 10.87, 2.76)
Building on it forms a hidden weak point.
Take an investor who buys on a rating, or a firm that hires on a credential. Each did what a careful person does and checked the certificate. What the certificate established was narrower than what they took from it, so B met the conditions B had read, and those were not the conditions that applied. What B holds is not false. It is held at a standard nobody established, and B builds on it. If it fails, the damage is in proportion to what was built on it, not to the size of the original gap. At scale, this is how a whole field comes to rest on a position none of its members ever tested. Alexandria calls this an induced pseudo-hinge. (Treatise 10.88, 2.136, 2.109)
A closure can outlast the relation that caused it.
B went toward some ground several times and found nothing there for them, because A held it. B formed a position of their own: commitment there produces nothing. It was accurate when formed. Then A leaves and the ground opens, but B’s position does not change, because the only thing that would test it is going there, and the position says not to.
It travels with B to every new role and place, and from inside it is indistinguishable from an accurate reading of ground that is still closed. Only two things move it: a commitment B will not make, or an arrival by someone who never held the position and so went where B would not. (Treatise 10.78, 2.86, 4.39)
Where every route meets, and the only part anyone else can see.
When B decides, three things were already set. The options were the ones the relation left in view. Each was weighed at the cost it had been made to appear. And the moment was set by how fast the ground was moving. All three arrive at one point in one person. So a decision that looks freely made tells you about A and B together, to the extent the relation set any of the three, and not about B alone.
The decision is also all that anyone else can see. What B holds and how B reads cannot be observed from outside. What B went toward and what B turned away from can be recorded, and nothing else can recover the relation. (Treatise 10.79, 10.77, 2.70)
What one person can do: the audit, and its limits.
One person can audit what they hold. B takes each position relied on and asks three questions. Did I work this out myself, or accept it from someone else? Will it bear the weight I am putting on it? And does it stand on its own support, or on the standing of whoever confirmed it? That last question reaches borrowed luminosity, and nothing else in the theory does.
Against the other routes the audit does nothing. It cannot reach what never returned, because there is nothing to examine. It cannot make an unreadable return readable. It cannot move what A stands on. And it cannot undo placement: knowing how you came to stand somewhere and standing somewhere else are different things. (Treatise 10.90, 10.91, 2.164)
What works is building, and only building.
Building the structure that makes returns readable answers illegibility. Building moves a vantage, because B stands where their confirmed structure puts them. Building capacity extends what a search can reach. All three are slow. Each is done by traversal, which nobody can do on B’s behalf, or by retrieval, and retrieval needs highway that someone has laid and left open. That is a disposition set at the collective level, not by the person who needs it. So the personal remedy runs back through what the collective leaves open. (Treatise 10.92, 2.67, 2.125)
Why these relations last without anyone defending them.
Closing something off is cheap. Withholding, declining to lay highway, standing where someone would have built, raising what a direction costs: each is done once and then stays done. Getting past it is expensive, because the only remedy is to build, building is done by traversal, and no one can traverse on another’s behalf. The effort to establish one of these relations and the effort to escape it are not comparable, and the gap does not close with time. (Treatise 10.93)