On the Severing of Power · Shi
论权力的凿削 · Shi

§I Introduction: Situating Severing Power

Power Theory is the dual series of the SAE Moral Law. The Moral Law works from the inside outward, treating the cases in which DD positions are symmetric or mutually sustaining; Power Theory works from the outside inward, treating the cases in which DD positions are asymmetric and not mutually sustaining. The Prequel cleared the referential space of power. Paper 1 established the 13DD point of origin and the three layers of the source, made the power-right duality explicit, and established the three theorems of remainder, duality, and reflexive-binding. Paper 2 unfolded the geometry, phase, and dimension of power along the three morphological axes, and staked the interface with the Means and Ends Kingdoms. Paper 3 cut apart the three positions of recognition and the state machine of operational-level recognition. Paper 4 unfolded the limit function of the remainder along T3 into a full domain. Paper 5 unfolded the developmental dynamics of power along T5 and carried severing to the trajectory level, that is, the right-orientation and rate of the slow and middle segments. The task of the present paper on this groundwork is single and determinate: to unfold the complete internal dynamics of severing power along the neutral operational definition of T1.

The present paper establishes no new T theorem. T1 was staked in Paper 1; Paper 6 unfolds along its neutral operational definition, depends one-directionally on Papers 1 through 5, and does not redefine them in return. Paper 5 §VIII explicitly left to this paper the internal anatomy of the locking action, the accumulation of T6 reflexive-binding under locking, the move of the one under power from end-position to means-position, and the equivalence of severing-mode power with the triggering of T6; Paper 2 §VIII staked the Means and Ends Kingdoms interface and left the equivalence of the M5 candidate criterion with T6 to this paper. The present paper discharges both commissions. It picks up the trajectory level already established by Paper 5 and descends to the locking mechanism itself; it does not redefine Paper 5's trajectories.

The central claim in one sentence: severing power is a self-consuming structure generated by cancelling the interrupt-right of the one under power. It buys internal stability with rigidification, and the bill for this trade is written, in its entirety and endogenously, against the structure itself; the account is running now, and it waits for no one to come collecting. This claim carries directional value. The first four papers are strictly neutral; Paper 5 is the first value-bearing paper of the series, and the present paper is the second: cultivation is benign power, severing is not. But the value attaches to the direction discriminated by T6, not to the remainder; the remainder itself remains without value (L1, Paper 4).

One thing must be said plainly at the outset, or the failure boundaries of the later sections will be misread. That the bill is written against the structure as a whole does not mean it is written against every node, and above all does not mean it is written against the apex. The parasitic equilibrium of §VI will show that the apex of a severing structure can end well while the structure's account is never cleared. The central claim speaks of the structure's account, not the apex's account.

Severing power is the paper of this series most easily read as a moral indictment of power. To pin it to structural theory, this section sets out four clarifications in advance; they run through the whole paper.

First, value attaches to direction, not to the remainder. The non-benign character of severing is a directional property discriminated by whether the T6 locking action is triggered; it is not a sin, and not the result of a negative value assigned to the remainder. The reflexive-binding that recurs throughout is a mechanical cost, not retribution. Paper 1 §8 has already fixed this: the dissolution of severing power is a mechanical fact, not moral retribution, and not a law of history. The remainder remains without value throughout (L1, Paper 4); this paper does not let the value carried by direction seep back into the remainder.

Second, end-position and means-position are structural operations within A's operational structure, not A's intent or psychology. Moving B to means-position presupposes no wish in A to use B; an A who executes locking need not think of B as a thing at all. The locking action itself places B structurally in means-position, independently of A's subjective states. This de-psychologizing discipline follows Paper 4 §II. Wherever a picture of extraction appears below, it is to be read by one standard structure: A attempts one hundred percent absolute locking and is blocked by the physics of ρ ≠ ∅ (the compression cannot reach bottom); the part of B's subjecthood that is forced to overflow is not a resource A has graciously preserved, but the system's residual heat that A cannot press down, and A is compelled to supply force permanently to suppress it. This is the mechanical cost of the severing structure, not A's calculation.

Third, the word severing holds two offices in this paper, and they must be kept apart. One is the generic sense: the asymmetric severing operation of T1, by which A, through asymmetric operation, alters, restricts, or directs B's available DD positions and paths of movement, with no direction presupposed. The other is the directional sense: severing power as the counterpart of cultivating power, which obtains only when the T6 locking action is triggered. This paper writes the directional one, but the generic operation is its substrate. The bare word severing defaults to the generic sense; wherever the directional pair is at issue, the text writes severing-mode or cultivating-mode (Paper 2 §I, the terminological ecology).

Fourth, the ineliminability of the maintaining action is a neutral universal of all power, not severing's own. Every power requires an ineliminable maintaining action, because ρ ≠ ∅: no demarcated finite range is self-sustaining for any subject, and the remainder keeps overflowing it. What discriminates the triggering of T6 is neither the presence of a maintaining action nor the size of its cost, but the object of the maintenance: only when what is maintained is that B shall not regain the interrupt-right does the maintaining action become locking in the sense of T6. The A of cultivating power likewise invests in maintenance, but does not cancel B's interrupt-right: B can leave, and one who stays keeps to a temporary scaffold oriented toward the dissolution of the positional difference, a scaffold that dissolves itself as B rises in level (C3, Paper 5). The A of severing power invests in maintenance in order to cancel B's interrupt-right permanently. There exist, besides, intermediate states that cancel the interrupt-right temporarily while oriented toward its return (education, administrative trusteeship, therapeutic restraint): the lock is temporary and self-liquidating, with a scheduled path of return. This paper does not unfold the intermediate states; it only marks their place. T6 is a mechanical criterion and presupposes no severing-cultivation dichotomy (Paper 1 §8: or other non-locking forms of power).

Proposition Preview

This paper establishes five criteria, unfolded along the neutral operational definition of T1, with no new theorem. The five differ in type; a one-sentence preview of each is given here, with its type and burden of proof, to give the reader a map of the criteria. The full statements are distributed across §II through §VI.

S1, the ontological action of severing (definition). The ontological action of severing power in the locking orientation is the cancellation, within A's operational structure, of B's interrupt-right: the move of B from end-position to means-position; means-position is a positive structural fact, not the negative print of missing cultivation. Full statement in §II.

S2, the equivalence of means-position and locking (theorem). A power form moves B to means-position if and only if it maintains an established downward locking of B, thereby triggering T6; the pivot of the equivalence is the ineliminability of the maintaining action that cancels the interrupt-right, not the cost of maintenance. Full statement in §III.

S3, the internal anatomy of the locking action and the accumulation of reflexive-binding (mechanism). Maintaining the ineliminable locking action compresses A's own remainder and binds it into reflexive-binding; reflexive-binding is at the same time the cohesion by which the severing structure persists. Full statement in §IV.

S4, vector conversion and rigidification (mechanism). The degree of freedom A releases by stepping outside the script, once its direction points at the downward locking of B, is release and locking in one and the same action; release is cast on the spot into the fuel of reflexive-binding. Rigidification is the same mechanism read along the time axis. Full statement in §V.

S5, the two failure boundaries of severing (model and corollaries). Severing is enclosed in a corridor that narrows along the two dimensions of locking depth and time, with the T5 boundary on one side and the T6 boundary on the other; the negativity is fully built in. Full statement in §VI.

These five, together with the conceptual machinery established in Papers 1 through 5, are the complete toolkit of this paper. No further tool is introduced.

---

§II The Ontological Action of Severing

What severing power does to the one under power at the ontological level, in the locking orientation, is given by S1.

> Criterion S1. The ontological action of severing. The ontological action of severing power in the locking orientation is the cancellation of the interrupt-right of B's own law within the A-B relation: the move of B, within A's operational structure, from end-position (where B's own law holds the standing to interrupt A's script) to means-position (where B's usable subjecthood is preserved, but B's own law is demoted to a managed variable that cannot interrupt A's script). Means-position is a positive structural fact, not the negative print of missing cultivation: it is composed of four components, namely output availability, the demotion of reasons, the recoding of refusal as explainable, and the absence of any appeal path for the remainder; it is a topological engineering by which A grafts a new purpose-interface onto the top of B's agency. Locking need not fix B's particular choices; once the return topology of B's choice space is fixed to flow back into A's purpose function, the interrupt-right is cancelled (return-map lock). Severing, equally, does not extinguish the remainder: the remainder is only compressed, and cannot be compressed to zero (ρ ≠ ∅, L1). It stands in reverse counterpart to C3, the ontological action of cultivation (level-raising plus the dissolution of a local grid, the interrupt-right preserved, the range self-dissolving): severing is downward locking plus the cancellation of the interrupt-right, with the range held down by A's continuous supply of force.

To make end-position and means-position precise is the starting point of all the work of this paper. End-position is not B mattering to A in A's heart; means-position is not B being harmed or stripped of choices. End-position is a structural standing: B's own law holds, within the A-B relation, the standing to interrupt A's script, that is, B can by its own law rewrite the purpose-setting of the relation. Means-position is the position after that standing has been cancelled: B's own law can appear only as a managed variable in A's script. It can be recorded, explained, placated, plugged into A's purposes; it cannot interrupt A's purpose-setting of the relation.

This definition is colder than treating a person as a means and colder than turning a person into a thing, and more accurate than both. Colder, because it appeals neither to B's experience nor to A's malice; more accurate, because it draws the line between severing and the exit of power. To turn B fully into a thing is to exit power: when A's operational structure no longer contains a position at which B is registered as a subject, what A exercises is not power but the processing of things (Paper 1 §6; in full at §VI). Severing power precisely cannot go that far. It needs B to retain enough subjecthood to execute, adapt, respond, carry, internalize. What it wants is B's usable subjecthood, not subjecthood as such. To cancel the interrupt-right while preserving usable subjecthood: that is the exact calibration of means-position.

Means-position is a positive print, not a negative one. If severing were read merely as the absence of cultivation, the reverse of level-raising plus grid-dissolution, Paper 6 would have no content of its own; it would be Paper 5 with a minus sign. Means-position contains what the cultivating side simply does not have. The ontological action of cultivation removes structure: it dismantles the local grid A has imposed on B, lifts B toward a higher level, and the relation runs toward the dissolution of the positional difference, withdrawing once done (C3, Paper 5). The ontological action of severing does not remove structure; it grafts structure: the loop of agency that B ran for B's own purposes is cut and re-welded onto A's purpose loop. B's machinery of agency keeps running at full speed; only its output end has been re-routed. This is a transplant of the purpose loop, not a reduction of capacity.

From this, means-position is composed of four positive components; together they characterize a complete position, not an absence. First, output availability: B's actions, skills, judgment, time, reputation, social position, and indeed B's locked state itself, are taken into A's output function; B has not vanished, B has been made into an output port. Second, the demotion of reasons: B's reasons can still be heard, but only as management information, risk signals, or efficiency parameters; they are no longer reasons that can change the purpose-setting of the relation. Third, the recoding of refusal as explainable: B's refusal is not registered as a subject's interruption of the script; it is recoded as misunderstanding, emotion, immaturity, or insufficient information. The refusal has not disappeared; the refusal has been rewritten. Fourth, no appeal path for the remainder: B's remainder is still there, can still leak at low amplitude, drift, err; but no structural path carries it back up to the layer where A's purposes are set. It cannot reach the rules themselves. Taken together, the exact sense of means-position is this: B is registered as a usable subject, and not registered as an interrupting subject.

The fourth component, no appeal path for the remainder, carries one further task in §III: it is the criterion that separates means-position from the intermediate states. A temporary lock with a scheduled path of return has, by that very fact, an appeal path; it fails the fourth component and does not constitute means-position. This is marked here; the detail is in §III.

Locking need not fix B's particular choices. This matters most for modern severing; otherwise severing reads as if it applied only to crude rule. An ancient severing locks B onto a single path: you may only do this. A higher-grade severing gives B a wealth of choices, may even leave B feeling highly free, while fixing the return topology of all the choices to flow back into A's purpose function. Locking therefore has four layers, each less like suppression than the last. Option locking: you may only do these. Path locking: you may do much, but only along these paths. Return locking: you choose your paths freely, but every outcome returns to A. Interpretation locking: you will interpret the returning-to-A as your own growth, your own choice, your own rationality. The higher the layer, the less the lock holds particular choices and the more it holds the return map of choosing, up to B's interpretive frame for B's own choosing. At every layer the criterion is one and the same: does B's own law still hold the standing to interrupt A's script? Under a return-map lock, however many free choices B makes inside the local net, the topological endpoints of all options converge by force on A's purpose function; at the meta-level of the relation B can neither reach nor change the rules themselves, and B's interrupt-right is cancelled at the level of return topology. Holding to this one line, that topological convergence is the zeroing of the interrupt-right, the soft severings of platforms, algorithms, modern organizations, consumer freedom, and identity markets all fall under one criterion, without stretching the word locking.

Severing, equally, does not extinguish the remainder. This sentence must be held to the end under L1 (Paper 4), or the duality with cultivation loses its balance. Severing does not extinguish B's remainder; it compresses B's remainder, and the remainder cannot be compressed to zero. Cultivation does not extinguish the remainder (C3: level-raising plus grid-dissolution, the absolute ontological ρ persisting); severing does not extinguish it either. Neither direction touches the ontological fact that the remainder cannot be extinguished; they differ only in orientation, the one running toward the amplification of B's remainder (the releasing orientation), the other toward its compression (the locking orientation). This is the exact sense of the reverse counterpart of S1 and C3: cultivation is level-raising plus the dissolution of the local grid, the interrupt-right preserved, the local range dissolving itself as B rises; severing is downward locking plus the cancellation of the interrupt-right, the locked range held down by A's continuous supply of force. One range shrinks by itself until it vanishes; one range must be pressed down by A without end.

Here an interface with the Moral Law series must be shown in its positive-negative pairing, but only marked, not merged. The First Theorem of Moral Law Paper 1 says: one cannot but recognize the concrete other as an end (the affirmative form of inner legislation). Severing, carried into the language of the Kingdoms, says its reverse: to cancel B's interrupt-right is to move B out of end-position (the negative form of outer force). The two propositions form a precise duality in the language of the Means and Ends Kingdoms; T5, as the shared spine of the two series, here receives its negative expression. But this paper speaks the mechanical language of Power Theory; the complete duality of the two series is reserved for a crossover paper. Only the interface is marked here.

---

§III The Equivalence of Means-Position and Locking

S1 gave the ontological action of severing in the language of the Kingdom of Ends (the move of B to means-position). T6 of Paper 1 gives the directional criterion of severing in the language of DD mechanics (the locking action triggers reflexive-binding). The two criteria stand independently: one was proposed as the M5 candidate in Paper 2 §VIII, the other established as a theorem in Paper 1 §8. This section proves them equivalent. This is the theorem-grade work at the core of this paper, named and left to it by Paper 2 §VIII. The section thereby discharges that commission: through the anchoring of this equivalence, M5 is promoted from Paper 2's candidate criterion to a standing predicate; this is not the introduction of a new tool, but the completion of an argument Paper 2 left open.

> Criterion S2. The equivalence of means-position and locking. A power form moves B to means-position (M5; means-position here taken in the sense of the permanent cancellation of the interrupt-right, that is, with no path of return) if and only if it maintains an established downward locking of B, thereby triggering T6 reflexive-binding (locking read from the start in the return-map sense of S1). The pivot of the equivalence is not the cost of maintenance but the ineliminability of the maintaining action that cancels the interrupt-right: the finite range demarcated for B is not self-sustaining against B's remainder (ρ ≠ ∅, the remainder keeps overflowing every finite range), so A must continuously execute an action that prevents B's own law from regaining the interrupt-right; that action is the locking (T6). Conversely, A's executing downward locking is A's preventing B's available DD position from rising; once the locking is established, B's own law can exist only as a managed variable in A's script, and that is means-position. M5 speaks the language of the Kingdom of Ends, T6 the language of DD mechanics; they are two independently established predicates. To prove their equivalence is to prove the two predicates co-extensive, not co-intensive (the two concepts remain distinct); hence no definitional duplication.

The argument runs along the line of the interrupt-right, not along cost. This choice is what makes the section stand, and why cost will not do must be said first. That severing pays to maintain its lock is true, but cost cannot serve as the pivot of an equivalence. Cultivation also pays: C4 (Paper 5) established that cultivation propagates through a network along low-reflexive-binding-cost paths, and low cost is not zero cost. If severing were judged by the size of its costs, what would be obtained is at most a difference of degree (severing's maintenance costs more), not an equivalence; built on cost, the equivalence would soften into a quantitative correlation, and a quantitative correlation cannot keep cultivation from biting in from the side. What this section proves is not that severing's maintaining action costs more, but that the maintaining action which cancels the interrupt-right is ineliminable. Ineliminability is a modal fact, not a quantity of degree. The action necessarily carries costs, but what the section pins is that the action cannot be omitted, not how much it costs; the mechanical anatomy of the cost belongs to §IV.

First, from means-position to locking. Suppose B has been moved to means-position. By S1 this means that B's own law has lost, within this relation, the standing to interrupt A's script and can appear only as a managed variable; and the loss is permanent, with no scheduled path of return. But B is a subject at 13DD or above; ρ ≠ ∅; B's own law does not vanish by being demoted. B's unbounded possibility as a 13DD subject keeps overflowing the finite range A has demarcated (L1). A therefore faces a structural fact: the cancelled interrupt-right has an inextinguishable source that keeps trying to regain it. For means-position to keep holding, A must continuously execute one action: preventing B's own law from regaining the interrupt-right. The action is ineliminable: it cannot be performed once and then released, because B's overflow does not stop; the demarcated finite range is not self-sustaining against B's remainder, it does not hold itself in place, it must be pressed down by A without end. And this continuous action of preventing B's own law from rising is precisely downward locking in the sense of T6: the attempt to fix B's DD position downward. Hence means-position triggers T6.

Next, from locking to means-position. Suppose A executes a downward locking of B: the attempt to fix B's available DD position downward and prevent its rise. Ask whether the locking is accomplished. If B's own law still retains the standing to interrupt A's purposes, that is, if B can still rewrite the purpose-setting of the relation by its own law, then the locking is not yet accomplished, for B can at any moment push the pressed-down DD position back up from inside. Once the locking is truly established, B's own law can exist only as a managed variable in A's script, no longer interrupting A's purpose-setting. And own law as nothing but a managed variable is exactly the means-position given by S1. Hence an established locking is means-position.

Here the object quantified by the equivalence is calibrated. Both sides of the biconditional are taken in the established state: on the M5 side, the establishment of means-position (the interrupt-right already permanently cancelled); on the T6 side, the establishment of the locking (B's own law already nothing but a managed variable). An unaccomplished locking attempt, in which B's own law is still interrupting, is outside the range of the biconditional; its maintaining action triggers reflexive-binding all the same, but that is the fact of T6 as an action-triggered criterion (Paper 1 §8), and it sits on a different level from the established-state equivalence of this section; the two do not yield counterexamples to each other. The domain quantified over is the subject domain of this series: subjects are purposive beings, and within the domain every established locking serves some output endpoint, either B's product or B's locked state as a signal addressed to others (§IV); a suppression with no output endpoint whatever does not constitute an operational structure in the sense of T1 and lies outside the domain.

One further declaration on how the word locking is read in this proof. The downward locking in both directions is read from the start in the return-map sense of S1: what is locked is B's available DD position within this relation and its return topology, not B's DD development in general. Under soft severing, B can accumulate capacities inside the net, can even raise B's general DD position; what is fixed is B's meta-level standing within this relation, that is, B's choice return-map is pinned to the shape that flows back into A's purpose function, and B can neither reach nor change the rules themselves (§II). The meta-level standing being fixed is the own law's being unable to interrupt the script, which is means-position; hence the equivalence holds for all four layers of locking as one, and is not confined to option-layer hard locking.

The two directions together: moving B to means-position and executing a downward locking that triggers T6 are two faces of one structural fact. The nature of this equivalence must be stated precisely, or it will be misread as definitional duplication, as if the two criteria had been declared one thing in advance and then pronounced equivalent. M5 and T6 are not two phrasings of one criterion. M5 speaks the language of the Kingdom of Ends: it speaks of B's position within A's operational structure (end-position or means-position). T6 speaks the language of DD mechanics: it speaks of the operation A performs on B's DD position (whether it is fixed downward) and of the consequence of that operation for A itself (reflexive-binding). The two languages translate into each other but do not reduce to each other (the register fixed by Paper 2 §VIII). What this section proves is that the two independent predicates are co-extensive: a power form satisfies M5 if and only if it satisfies T6; on every concrete form the two are true together and false together. Co-extension is not co-intension: equilateral triangle and equiangular triangle are co-extensive, picking out exactly the same class of figures, yet they are two concepts, one about sides, one about angles. So with M5 and T6: they pick out exactly the same class of power forms, yet they are two concepts, one about position, one about mechanics. The section upgrades the inter-translatability of Paper 2 §VIII to equivalence while holding to non-reduction.

This equivalence proof must seal two openings, or its range will sweep across criteria already established upstream. The two seals stand side by side.

First, the seal toward C3, cultivation. The fourth clarification of §I has established that the ineliminability of the maintaining action is a neutral universal of all power; cultivation too has an ineliminable maintaining action. Unsealed, the proof above would be read as: whatever maintains ineliminably triggers T6; cultivation maintains, so cultivation triggers T6, in direct collision with C3 (cultivation does not trigger T6). The seal lies in the criterion itself: what the proof pins is not the universal that maintenance is ineliminable, but that the maintaining action which cancels the interrupt-right is ineliminable. The A of cultivation invests in maintenance, but what it maintains is a temporary scaffold oriented toward the dissolution of the positional difference; it does not cancel B's interrupt-right (B can leave, and one who stays keeps to a set of rules that runs toward its own dissolution). Cultivation's maintaining action is therefore not locking in the sense of T6 and triggers no reflexive-binding, consistent with C3. The discrimination of T6 lies always in whether the interrupt-right is cancelled, never in the presence of maintenance.

Second, the seal toward the intermediate states. §I has marked the intermediate states that cancel the interrupt-right temporarily while oriented toward its return (education, administrative trusteeship, therapeutic restraint). Read literally, a teacher who temporarily cancels a student's standing to rewrite the script seems to satisfy the left side of M5 (B's own law cannot, at this moment, interrupt the script), and the equivalence would sweep the intermediate states into T6, misjudging them as severing. The seal takes its path through the fourth component of S1 (no appeal path for the remainder) and through the permanence written into the M5 predicate. An intermediate state has a scheduled path of return, hence an appeal path, hence fails the fourth component of means-position; correspondingly, the left side of M5 takes the permanent cancellation of the interrupt-right (no path of return), which the temporary cancellation of an intermediate state does not satisfy. So the M5 side is sealed against the intermediate states: a teacher's temporary cancellation of a student's interrupt-right does not satisfy M5. The T6 side is false as well, and has its own mechanism, which must be written out: the maintaining action of an intermediate state does not point in the direction of downward fixing. The range of severing is held down by A's continuous supply of force (the counterpart clause of S1); the range of an intermediate state liquidates itself toward the scheduled return, and the action that maintains it works at undoing the fixing, not at fixing. A temporary restraint therefore does not constitute an established locking in the sense of T6 either. Both sides false, the biconditional is sealed against the intermediate states in both directions.

With this, the positional characterization of severing-mode power (M5, the move of B to means-position) and its mechanical criterion (T6, locking triggering reflexive-binding) are anchored as one and the same class of forms. The next section dissects the inside of that locking action: how maintaining the ineliminable action compresses A's own remainder and binds it into reflexive-binding.

---

§IV The Internal Anatomy of the Locking Action and the Accumulation of Reflexive-Binding

§III has proved that the action which maintains means-position is ineliminable. This section dissects the inside of that ineliminable action on the power-holder's side: how it compresses A's own remainder and binds it into reflexive-binding, how the binding accumulates under locking, and why the binding is at the same time the cohesion by which the severing structure persists. This is given by S3.

> Criterion S3. The internal anatomy of the locking action and the accumulation of reflexive-binding. Maintaining the ineliminable locking action that cancels the interrupt-right compresses the power-holder A's own remainder: A continuously invests operating room in holding down the overflow of B's remainder, and this holding-down presses A's possibility into the narrow channel of maintenance; A's remainder is compressed and bound into reflexive-binding (T6, L4, accumulating only under locking). The ρ layer of reflexive-binding is the locked form of A's remainder (inextinguishable, it cannot be compressed to zero); its structural-layer cost is A's adjustable freedom cast into maintenance tasks (exhaustible, it can be pressed through a threshold). Reflexive-binding is at the same time the cohesion by which the severing structure persists: it welds A into the very structure A has locked; A loses the freedom to withdraw, to change the script, to restore B's interrupt-right, and at the limit is locked together with B; T6 is therefore not only a cost but the endogenous mechanism by which severing stands as a structure. The cost of reflexive-binding accumulates and accelerates under locking, with no exponential or any other particular functional form specified, and no quantification. For locking exercised through an apparatus, the reflexive-binding lands on the 13DD key-node set that invokes and manages the apparatus, not on the apparatus (L4, P2 §V.5, non-dilutability).

First the composition of reflexive-binding must be stated clearly; the whole paper reads by this sentence. Reflexive-binding has two layers. Its ρ layer is the locked form of A's remainder: A's remainder as a 13DD subject (L1, symmetric across both sides of the relation) is, while the lock is maintained, compressed and bound into the maintaining action; this remainder does not die, it can no more be compressed to zero than B's, and reflexive-binding is a form of A's remainder, not its ashes. Its structural-layer cost is A's adjustable freedom cast into maintenance tasks: the possibilities of action A could have deployed elsewhere are taken up, inch by inch, as continuous supply of force against B's overflow, and this adjustable freedom is exhaustible; it can be pressed through a threshold (§VI). The two layers stand together: the remainder as ontological differential cannot be extinguished; the adjustable freedom as structural-layer margin of action can be used up. Wherever the accumulation of reflexive-binding is spoken of below, what accumulates is the share of adjustable freedom cast into maintenance tasks, not any diminution of the remainder itself.

The anatomy starts from the conclusion of §III. The finite range demarcated for B is not self-sustaining against B's remainder: ρ ≠ ∅ (L1), B's possibility keeps overflowing every finite range, and by L3 the overflow cannot be absorbed without loss. A's maintaining action is therefore not a one-time installation but a continuous holding-down: A must keep investing its own operating room in the one task of preventing B's own law from regaining the interrupt-right. And A is itself a 13DD subject; A's action can occur only through the finite grid of action the structure affords it (the symmetric side of L1). A binds an ever larger part of this already finite grid to the single use of maintaining the lock, and A's own remainder is compressed accordingly: A's possibility is pressed into the narrow channel of maintenance, and the out-of-script remainder that could unfold freely is ever more suppressed. This is the mechanism by which T6 occurs inside the locking action, and it runs in exactly the direction of Paper 1 §8: the deeper the locking, the more the power-holder's own remainder of subjecthood is compressed, until at the limit the power-holder and the one under power are locked into the same structure. L4 has already cut this compression apart from A's remainder as such: A's remainder exists regardless (it is there without any lock); reflexive-binding accumulates only under locking; reflexive-binding is what locking does to A's remainder, not A's remainder itself.

Reflexive-binding accumulates, and accelerates. It accumulates because the maintaining action is ineliminable and B's overflow does not stop: every beat of maintenance casts a new share of adjustable freedom into the maintenance tasks, and the irreversibility of the casting follows the non-Markov state machine of Paper 3 (a retraction is not a true undo): arrangements of action already bound into the structure are themselves a new piece of work to unbind. It accelerates because the deeper the lock, the more tightly B's remainder is pressed, the greater the pressure of overflow (the cumulative transition of L2's three states), the more must be held down per unit of time, the more freedom must be cast in. This paper does not specify the accumulation as exponential or as any particular functional form, and does not quantify it; accumulation-with-acceleration is a structural assertion, the functional form an empirical question, a discipline the series has held since Paper 2.

Up to here reflexive-binding is a cost. But to read it only as a cost is to miss its other face toward the severing structure, and that face answers a question S1 and S2 have not answered: why the severing structure stands at all. Read the same mechanism in the other direction: that A's adjustable freedom is cast into maintenance tasks means A has also lost the freedom to leave this structure. A cannot easily withdraw, for withdrawal means the maintaining action stops, the lock loosens, B's own law regains the interrupt-right, the purpose-interface A grafted onto B's agency loses current, and everything cast in so far is written off. A cannot easily change the script, for every joint of the script is already meshed with the arrangements of maintenance. A cannot even easily restore B's interrupt-right, for restoration is the admission that everything cast in was sunk. Reflexive-binding welds A into the structure. A deeply severing structure is therefore unusually stable, not because no one wants to leave, but because even the apex is welded in place by its own maintaining action. Reflexive-binding turns from a cost item into the endogenous cohesion by which the severing structure persists: the structure persists by trapping itself. This runs in the same direction as the non-dilutability of L4: reflexive-binding does not move and does not spread thin; it grows on every node that executes a maintaining action and welds each node to its own position. This is the ground on which severing stands as a structure, borrowing nothing from morality and nothing from the future. This cohesion writes why the structure does not fall apart now; the same mechanism has another reading along the time axis, why the structure cannot change for the next moment, left to §V.

The ineliminable maintaining action also opens three internal accounts for A that are on the books now; none of them depends on any resistance from B, and none on any future event. The first: the inheritance of prediction. In cancelling B's interrupt-right A has cancelled B's standing to handle B's own future uncertainty by B's own law; but the uncertainty does not vanish with the standing; it is transferred to A. A must predict on B's behalf: where B will drift, when B will err, how B stays usable, how B does not cross the line. Severing does not reduce complexity; it moves B's adaptive complexity into A's maintenance system. The second: the debt of verification. As long as B remains a subject (and severing must keep B a subject, §II), A can never confirm once and for all that B has become a tool. Whether B's compliance is a deformation, whether B's execution carries a shadow topology (the misreading of Paper 4 §IV), A must verify again and again. Paper 4 has already booked this as the real friction misreading exacts from the structure (verification overhead); this paper shows it as severing's internal account: locking is not a one-time action but an institution of continuous verification. The third: the consumption of object quality. This account is given by the form, not by motive. The grafted interface of S1 plugs B's agentic output into A's purpose function; this is a structural fact of the severing form itself, independent of A's subjective needs; and the currency of the output is wide: labor, judgment, creation, compliance, and B's locked state itself. The last of these is the authority-display form: A takes no product from B; what is plugged into the output function is the state of B's being held down, sent as a signal toward the recognition network, its receiving end the bystander position within the network (Paper 3 §VI.3); through this signal A maintains power over others, taking the product and compliance of others. In this form the maintaining action is equally ineliminable: the legibility of the signal depends on the presence of the lock, and B's remainder keeps overflowing, so A must keep pressing; and since its output endpoint falls entirely on the recognition network, it is the severing most deeply dependent on the network. If a suppression takes neither B's product, nor B's locked state as a signal addressed to any other, nor serves the maintenance of power over any other, then it has no purpose function to speak of, does not constitute an operational structure in the sense of T1, and lies outside the domain of this series: the premise this series shares with the Moral Law is that every subject is a purposive being. And whatever the currency, what the severing form depresses is precisely the source of the currency: the interface wants B able to adapt, but not to set its own direction; wants B able to judge, but not to interrupt the script; the form uses B's products of subjecthood while weakening the subjective source of those products. This account has a further consequence in time, in §VI.

For locking exercised through an apparatus, where the reflexive-binding lands has been fixed by Paper 4 L4 and Paper 2 §V.5, and this paper invokes it: the binding lands on the 13DD key-node set that invokes and manages the apparatus, not on the apparatus; the apparatus has no subjecthood ρ and cannot bear reflexive-binding; the apparatus, as structure, only sediments structural-layer remainder (mediator cost). Nor is the binding diluted by headcount: every 13DD subject who takes part in the core action of the severing chain bears, within its own field, one complete share of T6. Apparatus-mediated division of labor cannot dilute T6; it can only replicate T6. This non-dilutability bears load again at the parasitic equilibrium and the consumable substrate of §VI.

Finally this section is pinned back onto the de-psychologizing discipline. In all of the anatomy above, A needs no particular psychological state. A need not hate B, need not want to use B, need not even be aware of locking. The ineliminability of the maintaining action is given by structure (the range is not self-sustaining); the accumulation of the binding is given by structure (the overflow does not stop); the cohesion is given by structure (the casting is irreversible). Likewise the picture of extraction is read by the standard structure of §I: the part of B's subjecthood forced to overflow is not a resource A has graciously preserved, but the system's residual heat A cannot press down; and the object A actively calibrates is B's usable subjecthood output (which, as a structural-layer product, can have functional value), not B's remainder (the remainder appears in this paper at exactly one argumentative position, that it cannot be compressed to zero, and never enters the register of value or calibration).

---

§V Vector Conversion and Rigidification

§IV has dissected how maintaining the lock casts A's adjustable freedom into maintenance tasks. This section looks at two further faces of that casting: how, within a single action, it co-occurs with release (vector conversion), and what it reads as along the time axis (rigidification). This is given by S4.

> Criterion S4. Vector conversion and rigidification. The degree of freedom A releases by stepping outside the script, once its direction points at the downward locking of B, is release and locking within one and the same action: the released freedom is cast on the spot into a chain on A's own body, the fuel of reflexive-binding (Paper 4 §VII). Reflexive-binding is therefore not a reaction that arrives after the locking; it is co-constituted with the locking action, and is the internal mechanism of severing's self-aggravation. Rigidification is the same locking mechanism (A's adjustable freedom cast into maintenance tasks) read along the time axis: the freedom cast in is no longer available for recombination, and when the time axis pushes in a new variable, A has no available freedom with which to change shape. Cohesion (§IV, why the structure does not fall apart now) and rigidification (why the structure cannot change for the next moment) are two readings of one mechanism, not two properties. This rigidification is the internal ground on which that branch of C2's exit bifurcation which holds the lock goes to physical collapse, taken up by Paper 7.

Vector conversion was staked in Paper 4 §VII; this paper unfolds it. The power-holder's out-of-script action is A's remainder expressing: A's possibility overflows the structural script, enters the transgressive state (L2), and shows as action the script has not encoded. Expression does not diminish the remainder (the remainder is an ontological differential, L1); the vehicle of the expression is A's adjustable freedom, the structural-layer margin of action, the exhaustible layer of §IV's composition sentence. Release itself is neutral; it can take infinitely many directions. But when the direction of this release points precisely at the downward fixing of B, release and locking coincide within one action. Paper 4's anchor is reused here: a jailer privately intensifying the treatment of a prisoner is at once his out-of-script action (the remainder expressing in the transgressive state) and a downward locking of the prisoner (pulling the T6 trigger). At that moment A casts the very share of adjustable freedom used by the expression into a chain on A's own body. Paper 4's staked phrase, that releasing the remainder casts the fuel of reflexive-binding, is read by this rule: fuel in the sense of casting feedstock, not in the sense of burning up; what is consumed is the adjustability of the vehicle, and what is cast is the cohesive building material that welds A into the structure (§IV), not the remainder burning away. Release is not freedom gained; when the direction of release is locking, to release is to cast the lock.

Vector conversion yields a key fact of timing: reflexive-binding is not a reaction that comes after the locking. It is not that there is first a lock, then an interval, and then the cost arrives. It is co-constituted with the locking action, in the same action, on the same beat. This co-occurrence is severing's internal mechanism of self-aggravation: every beat of maintenance is, in the same beat, a beat of fuel added to its own binding; there exists no period of locking without being bound. Severing therefore needs no external event to begin paying; it pays from the first locking action onward. This is the precise sense of the central claim's "the account is running now."

The same mechanism, read along the time axis, is rigidification. §IV has established that the adjustable freedom cast into maintenance tasks is no longer available for other uses; now follow the time axis and ask what that unavailability means. A structure meets a new situation by recombining itself: by drawing on freedom not yet bound, to generate new arrangements of response. A severing structure's adjustable freedom is cast, beat by beat, into maintenance tasks; the stock available for recombination falls beat by beat; when the time axis pushes in a new variable, the structure has no available freedom with which to change shape. It is not that it does not want to change; there is nothing left to change with. Here the composition sentence of §IV must be held: the remainder itself is not burned (it cannot be compressed to zero; L1 holds for A as well); what is burned into maintenance tasks is A's adjustable freedom, and what is locked shut on B's side are B's paths of remainder expression, not B's remainder burned. What rigidifies is the structure's capacity to recombine, not either side's remainder extinguished.

Cohesion and rigidification are two readings of one mechanism, not two properties. The cohesion of §IV speaks the statics of space: why the structure does not fall apart now; because every node's freedom has been used as building material, welded in place, presenting static rigidity toward internal perturbation. The rigidification of this section speaks the dynamics of time: why the structure cannot change for the next moment; because the same batch of freedom used as building material has solidified, and there is nothing to recombine when the new variable pushes in. One solidification, two readings: toward the inside it is stability; toward the next moment it is brittleness. A severing structure buys internal stability with rigidification; these are the two faces of one trade, not a stable structure that later went stiff.

This rigidification is the internal ground of the collapse branch of C2's exit bifurcation. Paper 5's C2 has established that the transition from the middle segment to the cultivating segment is a bifurcation at structural overload: one branch releases the hard lock at the thinnest point of stress and falls into the releasing orientation; the other holds the lock and goes to physical collapse, handed to Paper 7. This section supplies the internal mechanism of why the holding branch collapses: to hold is for the maintaining action to continue; for it to continue is for the casting to continue; for the casting to continue is for the recombinable freedom to keep falling; and §VI will show that the bearing of reflexive-binding has a reachable threshold. Paper 6 dissects to this point: the internal mechanism of self-aggravation on the locking side is installed; collapse as an event is left to Paper 7.

---

§VI The Two Failure Boundaries of Severing

The maintenance of severing keeps internal accounts (§IV), and aggravates itself beat by beat (§V). This section gathers the two into one model: the viable region of severing is a narrowing corridor with a failure boundary at either end. This is given by S5. The section also handles the two challenges sharpest against the framework of this paper, severing that appears stable and not self-defeating.

> Criterion S5. The two failure boundaries of severing. With locking depth as the axis (the deeper the lock, the more B's usable subjecthood output is pressed and the heavier A's maintenance load), the two failure boundaries of severing are called the T5 boundary and the T6 boundary. The viable region of severing is a corridor across the two dimensions of locking depth and time, enclosed by two descending failure frontiers: the extraction-death frontier gives, for each moment, the ceiling of viable depth, and it descends with sustained compression time (B's extractable output shows a decaying trend under compression, and the deeper the press the faster the source is eaten); the collapse frontier gives, for each depth, the ceiling of sustainable time, and it shortens with locking depth (the deeper the lock, the more adjustable freedom must be cast into maintenance tasks per unit of time). On any trajectory of sustained deepening or sustained maintenance both frontiers close in and the corridor narrows; no depth grants indefinite viability, shallow severing only narrows more slowly, hence optimal severing is shallow severing (optimal referring only to the depth position that walks the corridor longest, a description of structural sustainability, not a prescription for action). The T5 boundary splits into three layers: remainder-zero is an unreachable asymptote (ρ ≠ ∅); exit is a registration change within A's operational structure, available at any time; extraction yield at zero is the reachable functional threshold, and severing actually dies there. The T6 boundary is A's callable freedom falling below the structural maintenance threshold (a functional threshold, reachable), not A's remainder reaching zero (asymptotic, unreachable). The negativity of severing is fully built in: it grows on the three independent structural lines of the sediment layers, the narrowing corridor, and the drained reserve, and needs no resister.

§VI.1 The Corridor

The extraction of severing depends on a precise balance. B must be locked deeply enough: the purposes re-routed, the agency at A's call, or the grafted purpose-interface carries no current. And B must be pressed no further than the functional threshold: still a subject, the usable subjecthood output fully extractable, or the object of extraction evaporates (§II). The two requirements point in opposite directions, and this opposition unfolds, across the two dimensions of locking depth and time, into a corridor.

The two walls of the corridor are two failure frontiers on the plane of depth by time, both descending, and both already given by the mechanisms of the preceding sections. One is the extraction-death frontier; it gives, for each moment, the ceiling of viable depth. The deeper the lock, the more B's usable subjecthood output is pressed, the thinner the extractable agency, so this ceiling is finite to begin with; and the third account of §IV (the consumption of object quality) accumulates along time: B's extractable output shows a decaying trend under sustained compression, and the decay is dose-dependent, the deeper the press, the faster the quality source is eaten (still with no functional form specified), so this ceiling descends with time, and descends faster where the press is deep. The other is the collapse frontier; it gives, for each depth, the ceiling of sustainable time. The longer the maintenance, the more adjustable freedom has been cast into maintenance tasks (§IV §V), the closer A stands to the structural maintenance threshold, so this ceiling too is finite; and the deeper the lock, the more tightly B's remainder is pressed, the greater the pressure of overflow, the more freedom must be cast in per unit of time (§IV's mechanism of acceleration), so this ceiling shortens with depth. The two frontiers together: on any trajectory of sustained deepening or sustained maintenance, the ceiling of viable depth is falling and the ceiling of sustainable time is closing, and the viable region of severing narrows on the plane of depth by time. This is a corridor, not a cliff: not a one-sided slope of the deeper the riskier, but a passage held between two frontiers that grows narrower as it is walked. And no depth grants indefinite viability: the corridor narrows on every sustained trajectory; the shallow end only narrows more slowly.

The extraction-death frontier holds for all the currencies widened in §IV, including the signal currency of the authority-display form. The quality source of the signal is B's legibility as a subject: the locked state constitutes a display of power only while the recognition network still reads B as a subject; when a thing is displayed, what is displayed is no longer power (§VI.2, second layer). Deepening the lock does strengthen the signal in the short run, but what it consumes is precisely the signal's quality source: B's legible subjecthood is pressed, with depth and sustained compression, toward the reading of a thing, and the signal ceases to signify; at the limit, what the display has left is only the capacity to process things, and that can be had without any maintained lock at all, so the lock's extraction yield likewise goes to zero. This is the same account of object-quality consumption transposed onto the signal currency, the quality source replaced by legibility, the dose-dependence as before. The signal currency's output endpoint, moreover, falls entirely on the recognition network (§IV), and the recognition of the bystander position can flip (Paper 3 §VI.3); it is also the most exposed to erosion from the network side. The authority-display form therefore does not escape the extraction-death frontier: its death path is not the thinning of output, but the signal's ceasing to signify.

From this follows a cold corollary: optimal severing is shallow severing. Optimal has a precise sense here: the depth position at which the corridor can be walked longest. Deep severing presses the object of extraction thinner and shortens the ceiling of sustainable time at once; it strangles its own corridor. Shallow severing stands far from both ceilings, narrows slowly, walks long, and extracts inefficiently. A severing structure is endogenously unable to maximize extraction and persistence at the same time; this incompatibility is not an external constraint but the direct reading of two frontiers descending together; and shallowness grants no exemption, it only walks longer in a corridor that is still narrowing. The nature of this sentence must be pinned at once: optimal refers only to the extremum position of structural sustainability, a description of the internal geometry of the severing machine, not a prescription for action, and not advice on the technique of severing. The series does not predict, encourage, or evaluate any concrete political event; this sentence stands under that firewall. The corridor runs in the same direction as the exit bifurcation of Paper 5's C2: C2's death-cross (the cost of reflexive-binding crossing the extractable resource at high DD) is the trajectory-level expression of this corridor narrowing to its end.

§VI.2 The T5 Boundary

At one end of the corridor stands the T5 boundary. It must be split into three layers, and the three must not be welded into one thing.

The first layer, remainder-zero, is an unreachable asymptote. The limit of locking depth is to compress B's remainder to zero, one hundred percent colonization: all of B's possibility converged by A's grid, with nothing overflowing. But ρ ≠ ∅ (L1); the remainder cannot be compressed to zero; and remainder-at-zero and B's no longer being a subject are two phrasings of one and the same thing (one in the language of the remainder, one in the language of T5). This is a one-way limit characterization of the compression route: it says where the route's end lies, not that the cancellation of registration must travel there. The unreachability is confined to the interior of the track, the track on which the A-B power relation is still maintained and B's usable subjecthood is still extracted through the lock; along that track, severing never reaches one hundred percent colonization. This is the ontological wall; its function is to say that the locking direction has no terminus, not to say where severing fails.

The second layer, exit, is a change at the registration level, available at any time. A can stop registering B as a subject: process B directly as a thing. At that moment what A exercises is no longer power but the processing of things; A has unilaterally exited the social power relation with B and demoted itself to a processor of things (Paper 1 §6). Exit is not hitting the corridor's wall; it is leaving the corridor through the door: it does not pass through the limit of the first layer, is not the result of locking depth approaching its end, but a change of registration state within A's operational structure, available at any depth. After exit there is no severing, because there is no power.

The third layer, extraction yield at zero, is the reachable functional threshold, and severing actually dies here. Far before the remainder reaches zero, while B is still a subject, the extraction can already be finished: rising depth and accumulating compression time (the two frontiers of §VI.1) press B's usable subjecthood output below the extractable level; the grafted interface still carries current, but nothing comes through from the other side. Severing dies economically; the power relation persists in name only. This layer is reachable, and structurally it is hit before the first: severing need not wait for B to be pressed into a thing; it dies, while B is still a subject, of there being nothing left to extract. With the three layers apart, the precise sense of the T5 boundary is: the ontological wall cannot be reached (an asymptote), the registration door can be walked out of at any time (exit), and what is actually hit first is the functional threshold (extraction death).

§VI.3 The T6 Boundary

At the other end of the corridor stands the T6 boundary. It too must be stated by reachability. The zeroing of A's remainder is not this boundary: ρ ≠ ∅ holds for A symmetrically (L1), A's remainder can no more be compressed to zero than B's, and to take the exhaustion of A's remainder as the collapse condition would repeat the first layer's difficulty: a boundary that cannot be reached cannot serve as a condition of failure. The T6 boundary is a functional threshold: A's callable freedom falls below the structural maintenance threshold. Here the dual-layer composition sentence of §IV bears the load: the ρ layer of reflexive-binding (the locked form of A's remainder) does not die, but its structural-layer cost (the adjustable freedom cast into maintenance tasks) is exhaustible. The accumulation of maintaining a deep lock casts the adjustable freedom of A (or of the lowest locker in a parasitic chain, §VI.4) beat by beat into maintenance tasks; when what remains no longer suffices to execute the structure's own maintaining action, the threshold is crossed and the structure collapses. At that point B has not been erased; B is still inside the relation, pressed near the T5 functional threshold, while the structure collapses of being unable to afford its own maintenance, and the relation persists until it shatters. Asymptotes cannot be reached; thresholds can be crossed; collapse therefore has a mechanism of occurrence. Collapse as an event, and the dissolution dynamics after it, go to Paper 7.

The two boundaries are now apart: on the T5 side, severing dies economically at the functional threshold, or walks out through the registration door; on the T6 side, A's adjustable freedom is pressed through the maintenance threshold and the structure collapses toward Paper 7. On one side the extraction dies first; on the other the maintenance fails first; the end of the narrowing corridor is whichever of the two failures arrives sooner.

§VI.4 Two Challenges of Stable Severing

For the framework of this paper to stand, it must meet head-on the sharpest class of counter-positions: severing that looks stable, self-consistent, not self-defeating. If such severing existed, the corridor and the boundaries above would be no more than a romanticized structural tragedy. This class has two real forms, handled here on the spot, and one pseudo-form, handled in §VI.5.

The first form: the parasitic equilibrium, the spatial transfer of reflexive-binding. A does not itself maintain the lock on B; A locks an intermediate layer C and makes maintaining the lock on B C's low-cost path of ascent within the structure. C bears the maintenance willingly, because the positional gain it takes from B compensates its reflexive-binding. The middle layers of bureaucracies, the agents of tax-farming, each tier's wardens over the tier below in systems of rank: this structural shape recurs. The apex A locks almost no B directly, seems to bear almost no reflexive-binding; the structure is stable on net. Can this form pierce the framework of this paper?

A weak version must first be negated, and the negation must be performed by this paper itself: the proposition that the severing individual will in the end be dragged down by reflexive-binding is false. In a parasitic equilibrium the apex A can end well; the weight of the binding never lands on him directly. Any version that writes severing's self-defeat as the evildoer's final reckoning is moral retribution dressed as structural theory, and it is empirically simply wrong. The central claim of this paper contains no such weak version: the bill is written against the structure as a whole, which is not the same as against the apex (§I).

The strong version stands, but the ledger must be kept per node, not per chain. The non-dilutability of L4 bears the load here: reflexive-binding lands on the node that executes the maintaining action and cannot be transferred. What, then, moves along the parasitic chain? The maintaining action itself: A delegates the action of locking B to C. But delegation does not extinguish the apex's maintaining action; it only exchanges locking B for locking C, and A's locking of C remains A's own and remains ineliminable (C's own law is equally inextinguishable, C's assigned position equally not self-sustaining); A accumulates its own reflexive-binding for the A-C lock as before, the single entry merely smaller than for locking B directly. Every link of the chain is likewise: each generates its own ineliminable maintaining action and accumulates its own reflexive-binding. Hence delegation does not reduce the reflexive-binding in the system; it multiplies its entries: the longer the chain, the more ineliminable maintaining actions in the system, the more complete entries of reflexive-binding. This is not one binding shared out along the chain, and no summable total of binding need be presupposed; every node that executes a maintaining action bears one complete share of T6, and apparatus-mediated division of labor cannot dilute T6, only replicate it (P2 §V.5). The parasitic equilibrium has not extinguished reflexive-binding; it has multiplied its entries, only making the apex's own entry lighter.

Two sediment layers of different ledger items therefore build up along the parasitic chain, and they must not be conflated. Compression sediments down the chain onto the B at the bottom: B is pressed, lock upon lock, toward the T5 functional threshold, the extractable yield trending to zero; B locks no one, and what sediments on B is compressed remainder, the colonization-side quantity. Reflexive-binding sediments on the lowest locker C nearest to B: it executes the direct maintaining action on B, carries the heaviest single entry, and has no lower tier left to delegate to. Both sediment layers sit at the bottom of the chain, their position pinned by the T5 functional threshold, with no further transfer possible. The persistence of the parasitic chain therefore has an explicit mechanical premise: the extractable yield from the B at the bottom must exceed the compensation requirement of the lowest locker C, or C's willing bearing loses its ground. The thickening of the sediment is precisely the erosion of this premise: the compression sediment pushes B toward the functional threshold, that is, toward the break-point of the compensation chain; the stability of the chain and the sharpening of its most unstable point are two faces of one process. When the break-point is reached, the compensation chain unlatches from the bottom up; the unlatching as an event belongs to Paper 7; what this paper holds is the gradient itself. The apex A ends well, but A never escapes the ledger: A's entry for locking C keeps running, merely smaller; and A has surrendered for it the direct channel of disposal over the bottom, so that between A and B lies a chain A can no longer command link by link. The parasitic equilibrium is thereby collected into the thesis itself: the most stable severing manufactures its own most unstable point at maximal sharpness; the more it succeeds, the longer the chain, the smoother the transfer, the thicker the twin sediments at the bottom of the chain, the nearer the break-point of the compensation chain. Severing's self-defeat does not wait on the time axis for some day of reckoning (that is Paper 7's event layer); it sits in the spatial distribution of the structure, present now, a sediment the structure itself produces and positions.

The second form: the consumable substrate, the temporal re-stocking of reflexive-binding. A does not lock the same B for long; A presses B at speed, pushes B to the T5 functional threshold before the binding accumulated on this one B turns fatal, expels B, and plugs in the next B, riding a high-throughput steady state on a vast 13DD substrate. The labor camp of high mortality and the algorithmic platform of fast churn are the anchors of this form. It looks as if re-stocking lets it escape the accumulation of the binding.

To handle this form, the ontology of the substrate must first be stated, or a misjudgment enters at the registration level of T5. The substrate consists of many 13DD subjects, each with its own field of subjecthood and its own gate of negation; but within this structure they all stand on the locked side and execute no maintaining action, so reflexive-binding does not land on them (it is a quantity of the locker's side). And the substrate as an aggregate is not a 13DD subject; an aggregate has no gate of negation, so the binding does not land on the aggregate either (L4: apparatus and aggregates bear no reflexive-binding). Where does it land? Trace upward by L4: on the 13DD key-node set that invokes and manages this high-throughput machine, those who design the metrics, sign off the parameters, hold the dispatch-compression rate and the cadence of replacement.

The machine's maintaining action has not vanished with the re-stocking; it has only changed object: a single B is expelled, and the lock presses on the newcomer's return map as before; keeping the machine at full load is a continuous and ineliminable locking. The parameter regime (dispatch-compression rates, replacement cadence and the like) is here the codified instrument that holds down the return map of every online B; it is not the object of the lock. The deviations of the parameters have as their substantive source the aggregate of the online B-population's remainder expression: every own law is overflowing, and the aggregate presses on the parameter regime without pause. What the key nodes suppress through the parameter regime is, from first to last, this living aggregate overflow; the source of T6 is always the living subject, never the dead parameter, and the reflexive-binding re-anchors to these nodes at every act of replacement. These nodes cast their adjustable freedom without pause into the pipeline of patching, replacing, and pressing; they are bound to the console, having lost the freedom to stop the machine. At the same time, every B pushed to the functional threshold and expelled is one unabsorbable event of sedimentation: the substrate's consumption is on the books now, the law of accumulation of L3 showing itself on the substrate. What the substrate bears is not T6 (the binding has no anchor point on it) but the structural-layer sediment and the substrate depletion left by serial replacement. The consumable model thereby closes within the present-tense mechanisms of Paper 6: the binding has not vanished, it re-anchors to the key nodes; the consumption has not vanished, it sediments in the substrate's ledger. As for the exhaustion of the substrate, the machine gnawing its own chassis, that is the event of this machine colliding with its own ledger, and it is left to Paper 7, together with the full dissolution dynamics of the apparatus layer.

§VI.5 A Pseudo-Counterposition

One further proposal claims a stable, non-self-defeating severing: lock A and B together into a minimally dimensioned interlocked grid, eat up the freedom on both sides, so that reflexive-binding no longer accumulates and the structure becomes a deadlock that neither loosens nor erupts. This proposal is seductive and must be met head-on; but it is a pseudo-counterposition and poses no challenge to the framework.

The reasoning runs as follows. If the system truly no longer accumulates reflexive-binding, then by the mechanism of §IV it no longer executes the suppressing action (the binding accumulates only while the maintaining action of the lock occurs). And as long as B is still registered as a subject, the overflow of B's remainder does not stop; that is precisely the content of ρ ≠ ∅: what can stop is only the suppressing action, never the overflow. There are then only two cases. In the first, B is no longer registered as a subject: that is exit (§VI.2, second layer); there is no power, hence no severing, hence no question of a stable severing. In the second, the suppression has been withdrawn while both sides retain subject positions, and the relation is kept at a low dissipation of reflexive-binding: that is no longer the locking orientation; it has entered the low-energy neighborhood of the releasing orientation, the low-reflexive-binding-dissipation configuration into which the hard lock relaxes (C2), and it belongs to the cultivating segment (Paper 5), not to this paper. Neither outcome satisfies the dynamical definition of severing. The so-called stable deadlock is either not power or not severing; it is not a stable solution of severing, but the vanishing of severing's conceptual boundary.

§VI.6 The Built-in Negativity, and Exogenous Loads

The position of this paper's negativity can now be collected. The negativity of severing is fully built in; it grows on three independent structural lines, and not one of them needs a resister. The first line, the sediment layers (a distribution in space): a severing structure necessarily produces the twin sediment layers it cannot itself absorb, their position pinned at the bottom of the chain by the functional threshold, present now (§VI.4). The second line, the narrowing of the corridor (frontiers descending): the extraction-death frontier and the collapse frontier descend together across the two dimensions of depth and time, a pure functional reading of two machines already established (T5 and T6) on the operating parameters of severing, requiring no one to do anything (§VI.1). The third line, the draining of the reserve (rigidification): the freedom available for recombination is cast, beat by beat, into maintenance tasks, and stability toward the inside is brittleness toward the next moment (§V). All three lines are severing dismantling severing. Not one introduces a victim, a resister, or a righteous counterparty. On all three lines B can be fully compliant throughout, can even actively seek the lock: facing an unbounded option space it cannot converge on its own, B finds in locking exactly the supply of convergence; severing, as a reducer of possibility, can hold structural attraction for B, and many Bs not only do not resist the lock but actively seek to be arranged, prescribed, exempted from choosing. Here the valuelessness of the remainder receives its coldest confirmation: the remainder does not automatically generate resistance; that is a mechanical fact; the remainder merely never reaches zero, and it pushes in no direction. The negativity is complete all the same, for it grows on none of B's reactions; it grows on the structure's operation upon itself.

Last comes the exogenous load on the time axis; it must be cut cleanly apart from ρ, and the division of labor with Paper 7 must be cut cleanly as well. On physical time, any population necessarily collides with wholly new external loads it was not built to register: new adversaries, plagues, technological shocks; these serve here only as the briefest anchors and enter no argument. The exogenous load and ρ are not of one source: ρ is the overflow at the inner interface of the operational structure, A's structure unable to exhaust B's emergence, the differential between the structure and its own possibility; the exogenous load is the impact at the outer interface, the whole structure unable to exhaust the randomness of the environment, the encounter between the structure and the unincorporated outside. Their ontological sources differ and their interfaces of stress are orthogonal; they are not to be registered as one quantity. Yet at one level of abstraction the two are isomorphic: both are presentations of a finite structure's inability to exhaust an open domain, the one an incomputable quantity overflowing from within, the other an incomputable quantity striking from without, their landing points distinct. Severing's relation to both is given by rigidification: severing locks shut B's paths of remainder expression and burns A's adjustable freedom into maintenance tasks, that is, it drains the structure's adaptive reserve, whose physical substrate is precisely the adjustable freedom not yet cast into maintenance. The draining of the reserve is a present structural fact of Paper 6; it borrows nothing from any future event, and it is happening now (§V). How a structure whose reserve has been drained fails to recombine when the inevitably arriving exogenous load strikes, and goes to dissolution, depends on the arrival of an external event and belongs to Paper 7. One sentence cuts the two papers apart: this paper proves that internal friction has already burned out the suspension; how a car without suspension shatters on a stone from outside is Paper 7's business.

---

§VII Series Interfaces

This paper unfolds the internal dynamics of severing power along T1 and does not unfold dissolution. The interfaces are given here.

The interface with Paper 7, Dissolution, lies at dissolution. §V and §VI have given the internal ground on which the branch that holds the lock goes to physical collapse: rigidification drains the capacity to recombine, and reflexive-binding presses the bearing node's adjustable freedom toward the maintenance threshold. Collapse as an event, and the dissolution dynamics after it, are taken up by Paper 7. Paper 7's dissolution is fed by two branches together: on the side of the one under power, backlash raises the cost of maintenance and spreads through the recognition network (the remainder side of M4, Paper 4 §V; the flip of bystander recognition, Paper 3 §VI.3); on the power-holder's side, reflexive-binding crushes the structure from within (this paper). The impact events of exogenous loads, the exhaustion of the consumable substrate gnawing the chassis, and the complete dissolution dynamics of apparatus-layer remainder (the division fixed in P2 §V.5) are all left to Paper 7.

The interface with the Kingdom of Ends of the Moral Law series was marked in §II: to cancel the interrupt-right is to move B out of end-position, the precise reverse of the First Theorem of Moral Law Paper 1 (one cannot but recognize the concrete other as an end); T5, the shared spine of the two series, receives there its negative display in the language of the Kingdoms. The complete duality of the two series is reserved for a crossover paper.

The interface with 16DD: the series stops at 15DD; at the 16DD state of mutual non-doubt the concept of power dissolves, and it is excluded.

The concrete forms of severing on the three axes of Paper 2 are distributed across the sections by the needs of this paper's argument, with no separate catalogue of forms: covert severing and apparatus severing fall to §IV (the attribution of reflexive-binding and non-dilutability); algorithmic power and soft severing fall to §II and §III (the return-map lock); the parasitic equilibrium and the consumable substrate fall to §VI. The morphological dynamics of severing that Paper 2 §IX.3 named and left to this paper are thereby discharged.

---

§VIII Conclusion

This paper has unfolded the internal dynamics of severing power along the neutral operational definition of T1, the asymmetric-operation theorem of power, establishing no new T theorem. The two commissions named by Paper 5 §VIII and Paper 2 §VIII are thereby discharged.

The results of the paper collect into five points. One: the ontological action of severing power is the cancellation of the interrupt-right of B's own law within the A-B relation, the move of B from end-position to means-position; means-position is a positive structural fact composed of four components, and what severing wants is usable subjecthood, not subjecthood as such; the lock holds the return map of B's choice space, and need not hold particular choices (S1). Two: moving B to means-position is equivalent to maintaining an established downward locking of B that triggers T6; the pivot of the equivalence is the ineliminability of the maintaining action that cancels the interrupt-right, not the cost of maintenance; the two criteria are co-extensive and not co-intensive, so the equivalence is no definitional duplication (S2). Three: maintaining that ineliminable action compresses A's remainder and binds it into reflexive-binding, whose ρ layer does not die while its structural-layer cost is exhaustible; reflexive-binding is at the same time the cohesion that welds A into the structure, and severing persists by trapping itself; for locking through an apparatus the binding lands on the 13DD key-node set and cannot be diluted (S3). Four: the freedom released out of the script is cast, the moment its direction is locking, into the fuel of reflexive-binding; binding and locking are co-constituted, and severing pays from the first beat; the same casting reads along the time axis as rigidification, and stability toward the inside is brittleness toward the next moment (S4). Five: severing is enclosed in a corridor that narrows across the two dimensions of locking depth and time, with the T5 boundary split into three layers at one end (severing actually dies at the functional threshold where extraction yield reaches zero) and the T6 boundary at the other (adjustable freedom pressed through the maintenance threshold, giving collapse a mechanism of occurrence); the most stable severing manufactures its own most unstable point at maximal sharpness, and the negativity is fully built into the three lines of the sediment layers, the narrowing corridor, and the drained reserve (S5).

Severing power is, by all of this, a precise machine. It does not turn a person into a thing (that is the exit of power); it refashions subjecthood into uninterruptible instrumentality. It buys internal stability with rigidification, and the bill of this trade, the twin sediment layers, the narrowing corridor, the drained reserve, is written endogenously and in its entirety against the structure as a whole (which is not the apex's account), and the account is running now, waiting for no one to come collecting. Cultivation is power, at the end of its run in the releasing orientation, letting itself go; severing is power welding itself shut; and to be welded shut is stability toward the inside and brittleness toward the next moment. The two are the opposed orientations of one and the same neutral operation, each with a complete dynamics of its own.

Value attaches to direction, not to the remainder. The non-benign character of severing is a directional property discriminated by T6; the remainder itself remains without value from beginning to end, cannot be extinguished, and pushes in no direction. Resistance and dissolution are reserved for Paper 7, Dissolution. The work of the present paper is complete; Paper 7, Dissolution, takes up the complete structural account of resistance and dissolution.

---

Acknowledgments

This series is the dual of the SAE Moral Law series, the latter developed jointly by the author and Zesi Chen (陈则思). The present paper benefited from the divergence round and the outline and body reviews of a four-AI collaborative research methodology (Zilu, Zigong, Zixia, Gongxihua). At the outline stage the four reviewers' revisions were incorporated, among them: the corridor model rewritten from single-axis inverse functions into the two dimensions of locking depth and time; the symmetric application of ρ ≠ ∅ to A's side (reflexive-binding rewritten as the locked form of A's remainder rather than its ashes, the collapse trigger as adjustable freedom falling below the maintenance threshold); the T5 boundary split into asymptote, exit, and functional threshold; and the parasitic-equilibrium ledger rebooked per node (what transfers along the chain is the maintaining action, not the binding; the twin sediment layers kept as separate ledger items). At the body stage, the equivalence proof was calibrated to the established state and read throughout in the return-map sense; the intermediate states were sealed in both directions through permanence and the direction of fixing; the corridor was rewritten as two descending failure frontiers; the composition of reflexive-binding was unified into the dual-layer sentence of ρ layer and structural layer; the binding of the consumable model was re-anchored to the living aggregate overflow of the online subjects; the consumption of object quality was rewritten from motive to form, with the currency of output widened to the locked state as signal; and the coverage of the signal currency on the extraction-death frontier was completed in the final review round. Gongxihua gave the final sign-off on both the outline and the body.

---

References

The following works of the SAE series are cited. All are archived on Zenodo, with the Chinese and English versions in the same record. The cited DOIs are concept DOIs, always pointing to the latest version.

Qin, H. (2026). What Power Is Not (SAE Power Theory Series, Prequel). Zenodo. https://doi.org/10.5281/zenodo.20368867

Qin, H. (2026). On the Origin of Power (SAE Power Theory Series, Paper 1). Zenodo. https://doi.org/10.5281/zenodo.20370225

Qin, H. (2026). On the Morphology of Power (SAE Power Theory Series, Paper 2). Zenodo. https://doi.org/10.5281/zenodo.20455221

Qin, H. (2026). On the Recognition of Power (SAE Power Theory Series, Paper 3). Zenodo. https://doi.org/10.5281/zenodo.20480135

Qin, H. (2026). On the Limits of Power (SAE Power Theory Series, Paper 4). Zenodo. https://doi.org/10.5281/zenodo.20532999

Qin, H. (2026). On the Development of Power (SAE Power Theory Series, Paper 5). Zenodo. https://doi.org/10.5281/zenodo.20582211

Qin, H. (2026). 非 · Negativa: On Negation Prior to Being (SAE Methodology, Paper 0). Zenodo. https://doi.org/10.5281/zenodo.19544619

Qin, H. (2026). Via Rho: The Way of the Remainder (SAE Methodology, Paper 00). Zenodo. https://doi.org/10.5281/zenodo.19657439

Qin, H. (2026). Four Foundational Theorems of the Moral Law · Dao (SAE Moral Law Series, Paper 1). Zenodo. https://doi.org/10.5281/zenodo.20011019

The final paper of the series, Paper 7 (Dissolution), will follow in turn, sharing the T6 boundary and the C2 exit bifurcation with the present paper as interface.

---

(The English version of this paper is an independent rewrite by Han Qin, 2026-06-11.)

§I 引言: 凿削性权力的定位

权力论是 SAE 道德律论的对偶系列。 道德律自内而外处理 DD 位置对称或互维持的情形, 权力论自外而内处理 DD 位置非对称且非互维持的情形。 前置篇清出了权力的指称空间, P1 立了 13DD 起源点与三层源, 显化了权力权利对偶, 立了余项, 对偶, 反缚三定理。 P2 沿形态学三轴展开了权力的几何, 相位, 维度, 并为目的王国与手段王国接口立桩。 P3 切清了承认的三个位格与操作层承认的状态机。 P4 沿 T3 把余项的限度功能展开为全域。 P5 沿 T5 把权力的发展动力学展开, 把凿削处理到轨迹层, 即慢段与中段的权利取向与发展速率。 本篇 P6 的任务在此地基上单一而确定: 沿 T1 权力非对称操作定理的中性定义把凿削性权力的完整内部动力学展开。

本篇不立新 T 定理。 T1 已在 P1 立桩, P6 沿 T1 的中性操作定义展开, 单向依赖 P1 P2 P3 P4 P5, 不反过来定义前篇。 P5 §VIII 已点名把凿削的锁定动作内部解剖, T6 反缚随锁定的累积, 被权力者由目的位移到手段位, 以及凿削性与 T6 触发的等价全部留给本篇; P2 §VIII 已为 Means 与 Ends Kingdom 接口立桩并把 M5 候选判据与 T6 的等价留给本篇。 本篇兑现这两处点名。 本篇接 P5 已立的轨迹层再下探锁定机制本身, 不重定义 P5 的轨迹。

核心断言一句话: 凿削性权力是取消被权力者中断权而生成的一种自耗结构。 它用僵化买得对内的稳定, 这笔交易的账单全部内生地记在结构自己头上, 此刻就记着, 不等任何人来讨。 这一断言带方向价值。 前四篇严格中性, P5 是系列中第一篇带方向价值的, 本篇是第二篇: 涵育是良性权力, 凿削不是。 但价值附在由 T6 判别的方向上, 不附在余项上; 余项本身仍无价 (P4 L1)。

这一断言还须先说破一处, 否则后文的失效边界会被误读。 账单记在结构整体头上, 不等于记在每一个节点头上, 尤其不等于记在顶端节点头上。 后文 §VI 的寄生平衡态会显出, 一个凿削结构的顶端可以善终而结构整体的账从未清零。 核心断言说的是结构整体的账, 不是顶端的账。

凿削性权力是本系列最容易被读作对权力的道德控诉的一篇。 为把它钉在结构论上, 本节前置四条廓清, 贯穿全篇。

其一, 价在方向, 不在余项。 凿削的非良性是由 T6 锁定动作是否触发判别的方向属性, 不是罪, 也不是余项被赋了负价的结果。 后文反复出现的反缚是力学成本, 不是报应。 P1 §8 已定: 凿削性权力的瓦解是力学事实, 不是道德报应, 也不是历史规律。 余项始终无价 (P4 L1), 本篇不让方向上的价值渗回余项。

其二, 目的位与手段位是 A 操作结构里的结构操作, 不是 A 的意图或心理。 把 B 移到手段位不预设 A 想利用 B; 一个执行锁定的 A 不需要在主观上想物化 B, 锁定动作本身就把 B 结构性地放在手段位, 与 A 的主观状态无关。 这一去心理学化纪律接 P4 §II。 后文凡涉及榨取的图景, 都按一个标准结构读取: A 试图执行百分之百的绝对锁定, 但遭遇 ρ ≠ ∅ 的阻击而压不到底, B 被迫溢出的那部分主体性不是 A 开恩保留的资源, 而是 A 压不下去的系统余热, A 被迫永久供力镇压这股余热。 这是凿削结构的力学代价, 不是 A 的算计。

其三, 凿削一词在本篇有两个职位, 须分清。 一是 generic 义, 即 T1 的非对称凿削操作 (severing): A 通过非对称操作改变, 限制或定向 B 的可用 DD 位与位移路径, 不预设方向。 二是 directional 义, 即与涵育性权力对举的凿削性权力 (severing power), 它只在 T6 锁定动作被触发时成立。 本篇写的是 directional 那个, 但 generic 操作是它的底材。 裸词凿削默认 generic 义; 提及方向对偶处, 明确写成凿削性或涵育性 (P2 §I 术语生态)。

其四, 维持动作不可省, 是一切权力共有的中性共相, 不是凿削独有的。 任何权力都需要不可省的维持动作, 因为 ρ ≠ ∅, 被划定的有限范围对任一主体都不自持, 余项持续溢出。 触发 T6 的判别不在维持动作的有无, 也不在维持成本的高低, 而在维持的对象: 只有当维持的对象是 B 不能重获中断权时, 维持动作才成为 T6 义的锁定。 涵育性权力的 A 同样投入维持, 但不取消 B 的中断权: B 可以走, 想留下则守一套朝位差消解而去的临时支撑, 这一支撑随 B 升层而自我消解 (P5 C3)。 凿削性权力的 A 投入维持, 是为永久取消 B 的中断权。 此外还有一类临时取消中断权但朝归还而去的中间态, 例如教育, 行政托管, 治疗性约束: 锁是临时的, 自我清算的, 有一条排定的归还路径。 本篇不展开中间态, 只标出它的位置; T6 是力学判据, 不预设凿削涵育二分 (P1 §8 的或其他非锁定权力形态)。

命题预告

本篇立五条判据, 沿 T1 的中性操作定义展开 directional 凿削性, 不立新 T。 五条类型不一, 此处先给一句话预告并标出各条的类型与证明义务, 给读者一张判据地图。 完整陈述分布于 §II 至 §VI。

S1 凿削的本体动作 (定义). 凿削性权力在锁定取向上的本体动作, 是在 A 的操作结构里取消 B 的中断权, 即把 B 从目的位移到手段位; 手段位是正向的结构事实, 不是缺涵育的负片。 完整陈述见 §II。

S2 手段位与锁定的等价 (定理). 把 B 移到手段位当且仅当对 B 执行向下锁定从而触发 T6; 等价的支点是取消中断权的维持动作不可省, 不是维持成本。 完整陈述见 §III。

S3 锁定动作的内部解剖与反缚累积 (机制解剖). 维持那个不可省的锁定动作把 A 自己的余项压缩并绑定为反缚; 反缚同时是凿削结构持存的黏合。 完整陈述见 §IV。

S4 矢量转化与僵化 (机制解剖). A 越脚本释放的自由度方向指向锁定 B 时, 在同一动作里既是释放又是锁定, 释放即铸成反缚的燃料; 僵化是同一机制在时间轴上的读数。 完整陈述见 §V。

S5 凿削的两个失效边界 (模型加推论). 凿削被夹在锁定深度与时间两维上收窄的走廊里, 一边是 T5 边界, 一边是 T6 边界, 否定性完全内置。 完整陈述见 §VI。

这五条加上 P1 至 P5 已立的概念机器, 构成本篇的全部工具。 本篇不再立新工具。

---

§II 凿削的本体动作

凿削性权力在锁定取向上对被权力者做的那一件本体层的事, 由 S1 给出。

> 判据 S1 凿削的本体动作. 凿削性权力 (severing power) 在锁定取向上的本体动作, 是取消被权力者 B 的 own law 在 A-B 关系中的中断权 (interrupt-right), 即在 A 的操作结构里把 B 从目的位 (own law 具有中断 A 脚本的资格) 移到手段位 (B 的可用主体性被保留, 但 own law 被降为不能中断 A 脚本的被管理变量)。 手段位是正向的结构事实, 不是缺涵育的负片: 它由输出可用性, 理由降格, 拒绝可解释化, 余项无上诉路径四个构件合成, 是 A 在 B 的能动性顶端嫁接一个新目的接口的拓扑工程。 锁定不必固定 B 的具体选择; 只要 B 的选择空间的返回拓扑被固定为回流 A 的目的函数, 中断权即被取消 (返回图锁定, return-map lock)。 凿削同样不灭余项: 余项只被压缩, 压不到零 (ρ ≠ ∅, L1)。 与 C3 涵育的本体动作 (升层加局域网格解除, 保留中断权, 范围自我消解) 反向对位, 凿削是向下锁定加中断权取消, 范围由 A 持续供力压住。

把目的位与手段位精确化, 是本篇全部工作的起点。 目的位不是 B 在 A 心里重要, 手段位也不是 B 被伤害或被剥夺了选择。 目的位是一个结构资格: B 的 own law 在 A-B 关系中具有中断 A 脚本的资格, 即 B 可以凭自己的法则改写这段关系的目的设定。 手段位是这个资格被取消之后的位置: B 的 own law 只能作为 A 脚本中的一个被管理变量出现, 它可以被记录, 被解释, 被安抚, 被接入 A 的目的, 但不能中断 A 对这段关系的目的设定。

这个定义比把人当手段或把人变成物都冷, 也都准。 它冷, 因为它不诉诸 B 的体验, 不预设 A 的恶意; 它准, 因为它划清了凿削与权力退场的界。 完全把 B 变成物会让权力退场: 当 A 的操作结构里不再有 B 作为可登记主体的位置, A 施的就不是权力, 是对物的处理 (P1 §6, 完整见 §VI)。 凿削性权力恰恰不能走到那一步。 它需要 B 保留足够的主体性来执行, 适配, 回应, 承担, 内化。 它要的是 B 的可用主体性, 不是主体性本身。 取消中断权而保留可用主体性, 这就是手段位的精确刻度。

手段位是正片, 不是负片。 若把凿削仅仅读作缺涵育, 即涵育那条升层加解网格的反面, 那 P6 就没有自有内容, 只是 P5 取一个负号。 手段位有涵育那边根本没有的东西。 涵育的本体动作是减少结构: 拆掉 A 加在 B 身上的局域网格, 把 B 托往更高层级, 关系朝位差消解而去, 做完即撤 (P5 C3)。 凿削的本体动作不是减结构, 是嫁接结构: B 原本为自己的目的而运转的能动回路被切断, 重新焊接到 A 的目的回路上。 B 的能动机器仍在全速运转, 只是它的输出端被改了道。 这是一次目的回路的移植, 不是一次能力的削减。

由此, 手段位由四个正向构件合成, 它们刻画的是一个完整的位置, 不是一处缺席。 第一, 输出可用性: B 的行动, 技能, 判断, 时间, 声誉, 社交位置, 乃至 B 的被锁状态本身, 被纳入 A 的输出函数, B 不是消失了, 而是被做成一个输出端口。 第二, 理由降格: B 的理由仍可被听见, 但只作为管理信息, 风险信号或效率参数被处理, 它不再是能改变这段关系目的设定的理由。 第三, 拒绝可解释化: B 的不被登记为对脚本的主体中断, 而被重新编码为误会, 情绪, 未成熟或信息不足, 拒绝没有消失, 拒绝被改写。 第四, 余项无上诉路径: B 的余项仍在, 仍可以低幅泄出, 漂移, 出错, 但没有一条结构路径把它送回 A 的目的设定层, 它够不到规则本身。 四个构件合起来, 手段位的精确含义是: B 被登记为可用主体, 但不被登记为可中断主体。

第四个构件余项无上诉路径, 在 §III 还要承担一项工作: 它是区分手段位与中间态的判据。 一个有排定归还路径的临时锁定, 即有上诉路径, 因此不满足第四构件, 不构成手段位。 此处先标出, 详见 §III。

锁定不必固定 B 的具体选择。 这一点对现代凿削尤为要紧, 否则凿削会被读成只适用于粗暴统治。 一种古老的凿削把 B 锁在单一路径上, 你只能这样做。 但更高阶的凿削给 B 大量选择, 甚至让 B 觉得高度自由, 而把所有选择的返回拓扑固定为回流 A 的目的函数。 锁定因此有四层, 逐层更不像压制。 选项锁定: 你只能做这些。 路径锁定: 你可以做很多, 但只能沿这些路径移动。 返回锁定: 你可以自由选择路径, 但所有结果都回到 A。 解释锁定: 你会把回到 A 解释成自己的成长, 选择或理性。 越往高层, 锁定越不锁具体选择, 越锁选择的返回图乃至 B 对自己正在选择的解释框架。 但无论哪一层, 判据是同一个: B 的 own law 是否仍有中断 A 脚本的资格。 在返回图锁定下, B 在局域网内无论作多少种自由选择, 其所有选项的拓扑终点都强制收敛于 A 的目的函数, 故 B 在这段关系的元层级上永远触不到也改不了规则本身, B 的中断权在返回拓扑这一层被取消。 抓住拓扑收敛即中断权清零这一条, 平台, 算法, 现代组织, 消费自由, 身份市场这类柔性凿削就都进了同一判据, 不靠扩大锁定一词的词义。

凿削同样不灭余项。 这一句须按 P4 L1 守到底, 否则它会和涵育成对偶失衡。 凿削不消灭 B 的余项, 它压缩 B 的余项, 而余项压不到零。 涵育不消灭余项 (P5 C3: 升层加解网格, 绝对本体 ρ 续存), 凿削也不消灭余项, 两个方向都不触动余项不可消灭这一本体事实; 它们的区别只在取向, 一个朝放大 B 的余项而去 (松开取向), 一个朝压缩 B 的余项而去 (锁定取向)。 这正是 S1 与 C3 反向对位的精确含义: 涵育是升层加局域网格解除, 保留中断权, 那个局域范围随 B 升层而自我消解; 凿削是向下锁定加中断权取消, 那个被锁的范围由 A 持续供力压住。 一个范围会自己缩小直到消失, 一个范围要 A 一直按着。

这里须显出本系列与道德律论的一处正反面接口, 但只标出, 不并入道德律。 道德律 P1 第一定理说不得不承认具体他者为目的 (内立法的肯定形式)。 凿削延展到目的王国语言时, 说的是它的反面: 取消 B 的中断权即把 B 移出目的位 (外受力的否定形式)。 两条命题在目的与手段王国的语言下构成精确对偶, T5 作为双系列共同脊柱在此得到负向的表达。 但本篇走权力论的力学语言, 处理的是非对称且非互维持的情形; 权力论与道德律论的完整对偶留交叉论文, 本篇只在此标出接口。

---

§III 手段位与锁定的等价

S1 用目的王国的语言给出了凿削的本体动作 (把 B 移到手段位)。 P1 的 T6 则用 DD 力学的语言给出了凿削的方向判据 (锁定动作触发反缚)。 这两个判据各自独立成立, 一个在 P2 §VIII 作为 M5 候选被提出, 一个在 P1 §8 立为定理。 本节论证它们等价。 这是本篇核心定理级的工作, P2 §VIII 已点名留给本篇。 本节同时兑现这一委托: 经本节的等价锚定, M5 由 P2 的候选判据升格为正式谓词; 这不是另立新工具, 而是完成 P2 留下的论证。

> 判据 S2 手段位与锁定的等价. 一个权力形态把 B 移到手段位 (M5; 此处手段位取永久取消中断权义, 即无归还路径) 当且仅当它维持一个成立的对 B 的向下锁定从而触发 T6 反缚 (锁定自始按 S1 返回图义读取)。 等价的支点不是维持成本, 是取消中断权的维持动作的不可省: B 被划定的有限范围对 B 的余项不自持 (ρ ≠ ∅, 余项持续溢出任何有限范围), 故 A 必须持续执行一个阻止 B 的 own law 重获中断权的动作, 此动作即锁定 (T6)。 反向, A 执行向下锁定即阻止 B 的可用 DD 位上移, 锁定一旦成立, B 的 own law 只能作为 A 脚本中的被管理变量存在, 此即手段位。 M5 用目的王国语言, T6 用 DD 力学语言, 两者是独立立的两个谓词; 证等价是证两谓词同外延, 不是同内涵 (两个概念仍分立), 故非定义重复。

等价的论证走中断权这一条线, 不走成本。 这一选择是本节成立的关键, 须先说明为何不走成本。 凿削维持锁定要付成本, 这是真的, 但成本不能作等价的支点。 涵育也付成本: P5 C4 已立, 涵育靠低反缚成本路径在网络中传导, 低成本不是零成本。 若以成本之高低论凿削, 所得至多是一个程度差 (凿削的维持成本更高), 不是一个等价; 把等价建在成本上, 它会软化成量的相关, 而量的相关挡不住涵育从侧面咬进来。 本节要证的不是凿削的维持动作成本更高, 而是取消中断权的维持动作不可省。 不可省是一个模态事实, 不是一个程度量。 动作必然附带成本, 但本节锁定的是动作非省不可, 不是成本有多高; 成本的力学解剖留给 §IV。

先证从手段位到锁定。 设 B 被移到手段位。 按 S1, 这意味着 B 的 own law 在这段关系里失去了中断 A 脚本的资格, 只能作为被管理变量出现, 且这一失去是永久的, 没有排定的归还路径。 但 B 是 13DD 以上的主体, ρ ≠ ∅, B 的 own law 不因被降格而消失; B 作为 13DD 主体的无限可能性持续溢出 A 划定的那个有限范围 (L1)。 于是 A 面对一个结构上的事实: 被取消的中断权, 有一个不灭的源头一直在尝试重新获得它。 要让手段位持续成立, A 必须持续执行一个动作, 即阻止 B 的 own law 重新获得中断权。 这个动作不可省: 它不可被一次性完成而后撤手, 因为 B 的溢出不停; 被划定的有限范围对 B 的余项不自持, 不会自己保持住, 它要 A 一直按着。 而这个持续阻止 B 的 own law 上移的动作, 正是 T6 意义上的向下锁定: 试图把 B 的 DD 位向下固化。 故手段位触发 T6。

再证从锁定到手段位。 设 A 对 B 执行向下锁定, 即试图把 B 的可用 DD 位向下固化, 阻止其上移。 考察这个锁定是否完成。 若 B 的 own law 仍保留中断 A 目的的资格, 即 B 仍可凭自己的法则改写这段关系的目的设定, 那么锁定就还没有完成, 因为 B 随时可以从内部把被压下去的 DD 位顶回来。 锁定一旦真正成立, B 的 own law 就只能作为 A 脚本中的被管理变量存在, 不能再中断 A 的目的设定。 而 own law 只能作被管理变量, 这正是 S1 给出的手段位。 故锁定成立即手段位。

此处把等价的量化对象定标。 双条件的两侧都取成立态: M5 侧是手段位的成立 (中断权已被永久取消), T6 侧是锁定的成立 (B 的 own law 已只能作被管理变量)。 一个未完成的锁定尝试, B 的 own law 仍在中断, 不在双条件的射程内; 其维持动作照样在触发反缚, 那是 T6 作为动作触发判据的事实 (P1 §8), 与本节的成立态等价分属两层, 互不构成反例。 等价的量化域是本系列的主体域: 主体是目的性的存在, 域内凡成立的锁定都服务于某个输出端点, 或是 B 的产出, 或是 B 的被锁状态作为面向他者的信号 (§IV); 无任何输出端点的压制不构成 T1 义的操作结构, 不在域内。

还须声明本证里锁定一词的读法。 两个方向中的向下锁定自始按 S1 的返回图义读取: 锁定的对象是 B 在这段关系中的可用 DD 位及其返回拓扑, 不是 B 的一般 DD 发展。 柔性凿削下, B 可以在网内积累能力, 甚至升高一般的 DD 位; 被固化的是 B 在这段关系里的元层级资格, 即 B 的选择返回图被钉在回流 A 目的函数的形状上, B 触不到也改不了规则本身 (§II)。 元层级资格被固化, 即 own law 不能中断脚本, 即手段位; 故等价对四层锁定一体成立, 不限于选项层的硬锁定。

两个方向合起来, 把 B 移到手段位与对 B 执行向下锁定触发 T6 是同一个结构事实的两面。 此处须说清这个等价的性质, 否则它会被误读为定义重复, 即把两个判据预先说成一回事再宣布它们等价。 M5 与 T6 不是同一个判据的两种说法。 M5 用的是目的王国的语言, 它说的是 B 在 A 操作结构里的位格 (目的位还是手段位)。 T6 用的是 DD 力学的语言, 它说的是 A 对 B 的 DD 位做的操作 (是否向下固化) 以及这一操作对 A 自身的后果 (反缚)。 两套语言可以对译, 但不互相还原 (P2 §VIII 已立此口径)。 本节证的是这两个独立谓词同外延, 即一个权力形态满足 M5 当且仅当它满足 T6, 两者在每一个具体形态上同真同假。 同外延不是同内涵: 等边三角形与等角三角形同外延, 它们刻画完全相同的一类图形, 却是两个不同的概念, 一个讲边一个讲角。 M5 与 T6 同此: 它们刻画完全相同的一类权力形态, 却是两个不同的概念, 一个讲位格一个讲力学。 本节把 P2 §VIII 的可对译升级为等价, 同时守住不互还原。

这个等价证须封两处口, 否则它的射程会扫到上游已立的判据。 两处封口并列于此。

第一处, 对 C3 涵育封口。 §I 第四廓清已立, 维持动作不可省是一切权力共有的中性共相, 涵育也有不可省的维持动作。 若不封这一处, 上面的等价证会被读成只要维持动作不可省就触发 T6, 而涵育也维持, 于是涵育也触发 T6, 这与 C3 (涵育不触发 T6) 直接相撞。 封口在判据本身: 等价证锁定的不是维持动作不可省这个共相, 而是取消中断权的维持动作不可省。 涵育的 A 投入维持, 但维持的是一套朝位差消解而去的临时支撑, 它不取消 B 的中断权 (B 可以走, 想留下守的那套规则朝消解而去), 故涵育的维持动作不是 T6 意义上的锁定, 不触发反缚, 与 C3 一致。 触发 T6 的判别始终在中断权是否被取消, 不在维持的有无。

第二处, 对中间态封口。 §I 已标出存在临时取消中断权但朝归还而去的中间态 (教育, 行政托管, 治疗性约束)。 若按字面读, 一个教师在课堂上临时取消学生改写脚本的资格, 似乎也满足 M5 的左侧 (B 的 own law 此刻不能中断脚本), 于是被等价证一并扫进 T6, 这会把中间态误判为凿削。 封口取道 S1 的第四构件 (余项无上诉路径) 与 M5 谓词中的永久性。 中间态有一条排定的归还路径, 即有上诉路径, 因此不满足手段位的第四构件; 相应地, M5 左侧取的是永久取消中断权 (无归还路径), 中间态的临时取消不满足这个永久性条件。 故 M5 侧对中间态封闭: 教师临时取消学生中断权不满足 M5。 T6 侧另有自己的机制: 中间态的维持动作, 其方向不是向下固化。 凿削的范围由 A 持续供力压住 (S1 对位句), 中间态的范围朝排定的归还自我清算而去, 维持它的动作在解除固化而非固化, 故临时约束也不构成 T6 义的成立锁定。 两侧同假, 双条件对中间态在两个方向上都封闭。

至此, 凿削性的位格刻画 (M5, 把 B 移到手段位) 与凿削性的力学判据 (T6, 锁定触发反缚) 被锚定为同一类形态。 下一节解剖这个锁定动作的内部: 维持那个不可省的动作如何把 A 自己的余项压缩并绑定为反缚。

---

§IV 锁定动作的内部解剖与反缚累积

§III 已证, 维持手段位的动作不可省。 本节解剖这个不可省的动作在施权者一侧的内部: 它如何把 A 自己的余项压缩并绑定为反缚, 反缚如何随锁定累积, 以及反缚为何同时是凿削结构持存的黏合。 这由 S3 给出。

> 判据 S3 锁定动作的内部解剖与反缚累积. 维持那个取消中断权的不可省锁定动作压缩施权者 A 自己的余项: A 持续投入运作空间顶住 B 的余项溢出, 这一顶住的动作把 A 的可能性压进维持的窄道, A 的余项被压缩并绑定为反缚 (T6, L4, 仅锁定时累积)。 反缚的 ρ 层是 A 余项的锁定形态 (不灭, 压不到零), 其构层代价是 A 的可调自由度被铸入维护任务 (可耗, 可穿阈值)。 反缚同时是凿削结构持存的黏合: 它把 A 焊进自己锁的结构, A 失去抽身, 改脚本, 恢复 B 中断权的自由度, 极限处与 B 同被锁定; T6 因此不只是成本, 也是凿削何以作为结构站得住的内生机制。 反缚成本随锁定累积并加速, 不指定为指数或任何特定函数形式, 也不定量。 经装置施加的锁定其反缚落在调用并管理装置的 13DD 关键节点集身上, 不落装置 (L4, P2 §V.5 不可稀释性)。

先把反缚的成分说清, 全篇按此读取。 反缚有两层。 它的 ρ 层是 A 余项的锁定形态: A 作为 13DD 主体的余项 (L1, 对关系两侧对称) 在维持锁定时被压缩并绑定进维持动作, 这一余项不灭, 与 B 的余项一样压不到零, 反缚是 A 余项的一种形态, 不是 A 余项的灰烬。 它的构层代价是 A 的可调自由度被铸入维护任务: A 本可调去他处的施为可能性, 被一寸一寸地占用为顶住 B 溢出的持续供力, 这部分可调自由度是可耗的, 可以被压穿阈值 (§VI)。 两层并立: 余项作为本体差值不可消灭, 可调自由度作为构层的施为余量可被耗尽。 后文凡言反缚的累积, 累积的是被铸入维护任务的可调自由度占比, 不是余项本身在减少。

解剖从 §III 的结论起步。 B 被划定的有限范围对 B 的余项不自持: ρ ≠ ∅ (L1), B 的可能性持续溢出任何有限范围, 且这一溢出按 L3 不可被无损吸收。 于是 A 的维持动作不是一次性的安装, 是一个持续的顶住: A 必须把自己的运作空间持续投入到阻止 B 的 own law 重获中断权这一件事上。 而 A 自己同样是一个 13DD 主体, A 的施为只能经由结构所予的有限施为网格发生 (L1 对称侧)。 A 把这张本就有限的施为网格的越来越大一部分, 绑定到维持锁定这一个用途上, A 自己的余项随之被压缩: A 的可能性被压进维持的窄道, 能自由展开的越脚本余项越受抑。 这就是 T6 在锁定动作内部的发生机制, 它与 P1 §8 的表述完全同向: 锁定程度越深, A 主体性余项被压缩越多, 极限处权力者与被权力者同被锁定于该结构。 L4 已把这一压缩与 A 的余项本身切分: A 余项恒在 (不锁也有), 反缚仅锁定时累积, 反缚是锁定对 A 余项做的压缩, 不是 A 余项本身。

反缚累积并加速。 累积, 因为维持动作不可省且 B 的溢出不停, 每一拍维持都把新的一份可调自由度铸入维护任务, 而铸入的不可逆性接 P3 的非马尔可夫状态机 (撤销不是真正的 undo): 已被绑进结构的施为安排, 解开它本身又是一笔新的动作。 加速, 因为锁定越深, B 的余项被压得越紧, 溢出压力越大 (L2 三态的累积转态), 每单位时间须顶住的溢出越多, 须铸入的自由度越多。 本篇不把这一累积指定为指数或任何特定函数形式, 也不定量; 累积并加速是结构断言, 函数形式是经验问题, 系列自 P2 起守此纪律。

至此反缚是成本。 但只把反缚读作成本, 会漏掉它对凿削结构的另一面作用, 而这一面恰恰回答了一个 S1 与 S2 都没有回答的问题: 凿削结构为什么站得住。 换一个方向读同一个机制: A 的可调自由度被铸入维护任务, 意味着 A 也失去了离开这个结构的自由度。 A 不能轻易抽身, 因为抽身意味着维持动作中止, 锁松, B 的 own law 重获中断权, A 嫁接在 B 能动性上的目的接口断电, 此前铸入的一切作废。 A 不能轻易改脚本, 因为脚本的每一处都已与维持安排咬合。 A 甚至不能轻易恢复 B 的中断权, 因为恢复即承认此前的铸入是沉没的。 反缚把 A 焊死在结构里。 一个深度凿削的结构因此异常稳定, 不是因为没有人想走, 而是因为连顶端都被自己的维持动作焊在原地。 反缚由成本项翻为凿削结构持存的内生黏合: 结构靠困住自己来存续。 这与 L4 的不可稀释性同向: 反缚不挪不摊, 它就长在每一个执行维持动作的节点身上, 把每一个节点都焊在各自的位置上。 这是凿削何以作为结构站得住的地基, 完全不借道德, 不借未来。 这一黏合写的是结构此刻为何散不了; 同一机制在时间轴上还有另一个读数, 即结构面对下一刻为何变不了, 留 §V。

维持动作不可省, 还给 A 带来三笔此刻就在记的内部账, 它们都不依赖 B 的任何反抗, 也不依赖任何未来事件。 第一笔, 预测继承。 A 取消了 B 的中断权, 即取消了 B 用自己的 own law 处理自身未来不确定性的资格; 但不确定性不因资格被取消而消失, 它转嫁给 A。 A 必须替 B 预测: B 会在哪里偏移, 何时出错, 怎样继续可用, 怎样不越界。 凿削不是减少复杂度, 是把 B 的自适应复杂度搬进 A 的维护系统。 第二笔, 校验债。 只要 B 仍是主体 (而凿削必须让 B 仍是主体, §II), A 就不能一次性确认 B 已经是工具。 B 的配合是不是变形, B 的执行是不是带着影子拓扑 (P4 §IV 的误读), A 必须反复校验。 P4 已把这笔校验成本记为误读向结构索取的真实摩擦 (verification overhead); 本篇把它显为凿削的内部账: 锁定不是一次性动作, 是一套持续校验的制度。 第三笔, 对象品质消耗。 这笔账由形态给出, 不由动机给出。 S1 的嫁接接口把 B 的能动输出接入 A 的目的函数, 这是凿削形态本身的结构事实, 与 A 的主观需求无关; 输出的通货是宽的, 劳动, 判断, 创造, 服从, 乃至 B 的被锁状态本身。 最后这种是立威形态: A 不取 B 的任何产出, 被接入输出函数的是 B 被压住这一状态, 它作为信号面向承认网络发送, 信号的接收端是网络中的旁观者位格 (P3 §VI.3), A 经这一信号维持对他者的权力, 取的是他者的产出与顺从。 此形态下维持动作同样不可省: 信号的可读性依赖锁的在场, 而 B 的余项溢出不停, A 必须不停按住; 且它的输出端点整个落在承认网络上, 是对网络依赖最深的一种凿削。 若一个压制既不取 B 的产出, 也不取 B 的被锁状态作为面向任何他者的信号, 也不服务于对任何他者的权力维持, 则它没有目的函数可言, 不构成 T1 义的操作结构, 不在本系列的讨论范畴: 本系列与道德律论共享的前提是, 每个主体都是目的性的存在。 而无论通货为何, 凿削形态压低的恰恰是通货的来源: 接口要 B 能适应, 但不许自定方向; 要 B 能判断, 但不许中断脚本; 形态利用 B 的主体性产物, 同时削弱这些产物的主体性来源。 这一笔账在 §VI 还有一个时间上的后果。

经由装置施加的锁定, 反缚落在哪里, P4 L4 与 P2 §V.5 已定, 本篇调用: 反缚落在调用并管理装置的 13DD 关键节点集身上, 不落装置; 装置无主体性 ρ, 不能承反缚, 装置作为构只沉积构层余项 (mediator cost)。 反缚不在人数上稀释: 每一个参与凿削链条核心动作的 13DD 主体, 在自身场域内承受一份完整的 T6。 装置化分工不能稀释 T6, 只能复制 T6。 这一不可稀释性在 §VI 的寄生平衡态与消耗件模型处还要承重。

最后把本节钉回去心理学化的纪律上。 上述全部解剖里, A 不需要任何特定的心理状态。 A 不需要恨 B, 不需要想利用 B, 甚至不需要意识到自己在锁定。 维持动作的不可省由结构给出 (范围不自持), 反缚的累积由结构给出 (溢出不停), 黏合由结构给出 (铸入不可逆)。 同样, 榨取图景按 §I 的标准结构读取: B 被迫溢出的那部分主体性不是 A 开恩保留的资源, 是 A 压不下去的系统余热; A 主动校准的对象是 B 的可用主体性输出 (它作为构层面的产出可以有功能价值), 不是 B 的余项 (余项在本篇只在压不到零这一个论证位出场, 永不进入价值与校准的语域)。

---

§V 矢量转化与僵化

§IV 解剖了维持锁定如何把 A 的可调自由度铸入维护任务。 本节看这一铸入的两个进一步面向: 它在单个动作内部如何与释放同构发生 (矢量转化), 以及它在时间轴上读出什么 (僵化)。 这由 S4 给出。

> 判据 S4 矢量转化与僵化. A 越出脚本释放的那份自由度, 一旦方向指向对 B 的向下锁定, 在同一个动作里既是释放又是锁定: 释放出的自由度当场铸成套在 A 自己身上的锁链, 成为反缚的燃料 (P4 §VII)。 反缚因此不是锁定之后才来的反应, 而是与锁定动作同构发生, 是凿削自我加重的内部机制。 僵化是同一锁定机制 (A 的可调自由度被铸入维护任务) 在时间轴上的读数: 被铸入的自由度不再可供重组, 时间轴推入新变量时, A 没有可用自由度去改形。 黏合 (§IV, 结构此刻为何散不了) 与僵化 (结构面对下一刻为何变不了) 是同一机制的两个读数, 不是两个性质。 此僵化是 C2 转出分叉里死扛锁定那一支走向物理崩塌的内部根据, 由 P7 接续。

矢量转化在 P4 §VII 已立桩, 本篇展开它。 施权者的越脚本施为是 A 的余项在表达: A 的可能性溢出结构脚本, 进入越界态 (L2), 显为脚本未编码的施为。 表达不减余项 (余项是本体差值, L1); 表达的载具是 A 的可调自由度, 即构层的施为余量, §IV 成分句里可耗的那一层。 释放本身是中性的, 它可以取无限种方向。 但当这一释放的方向恰好指向对 B 的向下固化时, 同一个动作里释放与锁定重合了。 P4 的锚点在此复用: 一个狱卒私自加重对囚犯的处置, 既是他越出脚本的施为 (余项在越界态表达), 也是一次对囚犯的向下锁定 (扣动 T6)。 此时 A 把表达所用的那份可调自由度, 当场铸成了套在自己身上的锁链。 P4 的立桩语释放余项即铸成反缚燃料按此读取: 燃料取铸造供料义, 不取燃尽义; 所耗的是载具的可调性, 所铸成的是把 A 焊进结构的黏合建材 (§IV), 不是余项本身在燃尽。 释放并不等于获得自由; 当释放的方向是锁定时, 释放即铸锁。

矢量转化给出反缚的一个关键时序事实: 反缚不是锁定之后才来的反应。 它不是先有锁定, 再过一段时间, 然后成本到账。 它与锁定动作同构发生, 在同一个动作里, 同一拍。 这一同构是凿削自我加重的内部机制: 凿削的每一拍维持, 同时就是给自身的反缚添的一拍燃料, 不存在一个只锁定而不被反缚的时段。 凿削因此不需要等待任何外部事件来开始支付, 它从第一个锁定动作起就在支付。 这是本篇核心断言里此刻就记着的精确含义。

同一个机制, 换到时间轴上读, 是僵化。 §IV 已立, 被铸入维护任务的可调自由度不再可供他用; 现在沿时间轴看这一不再可供他用意味着什么。 一个结构应对新局面, 靠的是重组自己: 动用尚未被绑定的自由度去生成新的应对安排。 凿削结构的可调自由度被一拍一拍铸入维护任务, 可供重组的存量逐拍下降; 时间轴推入新变量时, 结构没有可用的自由度去改形。 它不是不想变, 是没有东西可拿来变。 这里须按 §IV 的成分句守住本体: 余项本身不被烧毁, 压不到零 (L1 对 A 同样成立); 被烧成维护任务的是 A 的可调自由度, 被锁住的是 B 的余项表达路径。 僵掉的是结构的重组能力, 不是哪一方的余项被消灭了。

黏合与僵化是同一机制的两个读数, 不是两个性质。 §IV 的黏合说的是空间上的静力学: 结构此刻为何散不了, 因为每个节点的自由度都被当成了建材, 焊死在各自位置, 对内部扰动呈现静态的刚性。 本节的僵化说的是时间上的动力学: 结构面对下一刻为何变不了, 因为同一批被当成建材的自由度已经凝固, 新变量推进来时无可重组。 一笔凝固, 两个读数: 对内, 它是稳定; 对下一刻, 它是脆性。 凿削结构用僵化买对内稳定, 这是同一笔交易的两面, 不是先有一个稳定的结构后来才变僵。

这一僵化是 C2 转出分叉的崩塌支的内部根据。 P5 C2 已立, 中段到涵育段的转出是结构过载处的分叉, 一支在应力最薄处松开硬锁定落入释放, 另一支死扛锁定走向物理崩塌, 交由 P7。 本节给出死扛那一支为什么走向崩塌的内部机制: 死扛即维持动作继续, 维持继续即铸入继续, 铸入继续即可供重组的自由度继续下降, 而反缚的承载在 §VI 将显出有一个可达的阈值。 P6 解剖到这里为止: 锁定侧的内部加重机制装好, 崩塌作为事件留 P7。

---

§VI 凿削的两个失效边界

凿削的维持有内部账 (§IV), 且逐拍自我加重 (§V)。 本节把这两条收进一个模型: 凿削的可行区是一条收窄的走廊, 两端各有一个失效边界。 这由 S5 给出。 本节同时处理两类对本篇框架最锋利的挑战, 即看起来稳定而不自败的凿削。

> 判据 S5 凿削的两个失效边界. 以锁定深度为轴 (锁定越深, B 的可用主体性输出越受压, A 的维持负荷越重), 凿削的两个失效边界称 T5 边界与 T6 边界。 凿削的可行区是锁定深度与时间两维上的一条走廊, 由两条下行的失效前沿围出: 榨取死亡前沿给出每一时刻的可行深度天花板, 它随持续压缩的时间下移 (B 的可榨能动性输出在压缩下呈衰减趋势, 且压得越深蚀得越快); 崩塌前沿给出每一深度的可维持时间天花板, 它随锁定深度收短 (锁得越深, 每单位时间须铸入维护任务的可调自由度越多)。 任何持续加深或持续维持的轨迹上, 两条前沿都在逼近, 走廊收窄; 没有任何深度给出无限期的可行, 浅凿削只是收得慢。 T5 边界拆三层: 余项归零是不可达渐近线 (ρ ≠ ∅); 退场是 A 操作结构里的登记变更, 随时可发生; 榨取增益归零是可达的功能阈值, 凿削实际死在这里。 T6 边界是 A 的可调用自由度跌破结构维持阈值 (功能阈值, 可达), 不是 A 余项归零 (渐近, 到不了)。 凿削的否定性完全内置: 它长在反缚沉积层, 走廊收窄, 储备抽干三条独立结构线上, 不需要任何反抗者。

§VI.1 走廊

凿削的榨取依赖一个精确的平衡。 B 必须被锁得足够深: 目的被改道, 能动性归 A 调用, 否则嫁接的目的接口不通电。 B 又必须被压得仍在功能阈值之上: 仍是主体, 可用主体性输出完整可榨, 否则榨取对象蒸发 (§II)。 两个要求方向相反, 这一相反在锁定深度与时间两维上展开为一条走廊。

走廊的两面墙是深度乘时间平面上的两条失效前沿, 都在下行, 它们都已由前文的机制直接给出。 一条是榨取死亡前沿, 它给出每一时刻的可行深度天花板。 锁定越深, B 的可用主体性输出越受压, 可榨的能动性越薄, 故这块天花板本就有限; 而 §IV 的第三笔账 (对象品质消耗) 沿时间积累, B 的可榨输出在持续压缩下呈衰减趋势, 且衰减带剂量性, 压得越深, 品质来源被蚀得越快 (仍不指定函数形式), 故这块天花板随时间下移, 深处下移得更快。 另一条是崩塌前沿, 它给出每一深度的可维持时间天花板。 维持越久, 被铸入维护任务的可调自由度越多 (§IV §V), A 离结构维持阈值越近, 故这块天花板同样有限; 而锁得越深, B 的余项被压得越紧, 溢出压力越大, 每单位时间须铸入的自由度越多 (§IV 的加速机制), 故这块天花板随深度收短。 两条前沿合起来: 任何持续加深或持续维持的轨迹上, 可行深度的天花板在降, 可维持时间的天花板在收, 凿削的可行区在深度乘时间的平面上收窄。 这是一条走廊, 不是一道悬崖: 不是越深越险的单边坡, 是两条前沿夹出的一条越走越窄的通道。 且没有任何深度给出无限期的可行: 走廊在每一条持续轨迹上都收窄, 浅处只是收得慢。

榨取死亡前沿对 §IV 放宽后的全部通货成立, 包括立威形态的信号通货。 信号的品质来源是 B 作为主体的可辨识性: 被锁状态只有在承认网络仍把 B 读作主体时才构成权力的展示, 物被展示时, 展示的不再是权力 (§VI.2 第二层)。 加深锁定在短期内确实增强信号, 但它消耗的正是信号的品质来源: B 的可辨识主体性随深度与持续压缩被压向物的读数, 信号随之失义; 极限处展示所剩的只是处理物的能力, 而那不再需要任何被维持的锁定即可取得, 锁的榨取增益同样归零。 这是对象品质消耗在信号通货上的同一笔账, 品质来源换为可辨识性, 剂量性同前。 信号通货的输出端点又整个落在承认网络上 (§IV), 旁观者位格的承认可翻转 (P3 §VI.3), 它对网络侧的侵蚀也最敞开。 立威形态因此不逃出榨取死亡前沿: 它的死亡路径不是产出变薄, 是信号失义。

由此得一个冷推论: 最优凿削是浅凿削。 最优在此有精确含义: 走廊里能走得最长的深度位置。 深凿削同时把榨取对象压得更薄, 把可维持时间的天花板收得更短, 它自掐走廊; 浅凿削离两块天花板都远, 收得慢, 走得长, 但榨取效率低。 凿削结构内生地无法同时最大化榨取与存续, 这一不可兼得不是外部约束, 是两条前沿同时下行的直接读数; 而浅也不给出豁免, 它只是在一条仍在收窄的走廊里走得久一些。 须立即钉住此句的性质: 最优只指结构可持续性的极值位置, 是对凿削机器内部几何的描述, 不是行动处方, 也不是凿削技术的建议。 本系列不预测, 不鼓励, 不评价任何具体政治事件, 此句同受此防线约束。 这条走廊与 P5 C2 的转出分叉同向: C2 的死亡交叉 (反缚成本与可榨取资源在高 DD 处交叉) 正是这条走廊收窄到尽头的轨迹层表述。

§VI.2 T5 边界

走廊的一端是 T5 边界。 它须拆成三层, 三层不可焊成同一件事。

第一层, 余项归零, 是不可达的渐近线。 锁定深度的极限是把 B 的余项压到零, 即百分之百的殖民: B 的全部可能性被 A 的网格收敛, 无一溢出。 但 ρ ≠ ∅ (L1), 余项压不到零; 余项压到零与 B 不再是主体是同一件事的两种说法 (一个用余项语言, 一个用 T5 语言), 此为压缩路线的单向极限刻画, 它说压缩的尽头在哪里, 不说取消登记必须走到这里。 这一不能到达限定在轨道的内部, 即仍维持 A-B 权力关系, 仍经锁定榨取 B 的可用主体性的那条轨道; 沿这条轨道, 凿削永远到不了百分之百的殖民。 这是本体层的墙, 它的功能是说明锁定的方向上没有终点, 不是说明凿削在哪里失效。

第二层, 退场, 是登记层的变更, 随时可发生。 A 可以不再把 B 登记为主体: 直接把 B 当物处理。 那一刻 A 施的不再是权力, 是对物的处理, A 单方面退出了与 B 的社会权力关系, 降格为物的处理者 (P1 §6)。 退场不是撞上走廊的墙, 是从门里走出走廊: 它不经第一层的极限, 不是锁定深度逼近尽头的结果, 是 A 操作结构里登记状态的一次变更, 在任何深度上都可以发生。 退场之后没有凿削, 因为没有权力。

第三层, 榨取增益归零, 是可达的功能阈值, 凿削实际死在这里。 在余项远未压到零, B 仍是主体的地方, 榨取就可能先到头了: 锁定深度上升与压缩时间累积 (§VI.1 的两条耦合) 把 B 的可用主体性输出压到可榨水平之下, 嫁接接口仍通电, 但接口那头已经送不出东西。 凿削在经济上死亡, 权力关系名存实亡。 这一层是可达的, 而且在结构上先于第一层被撞上: 凿削不必等 B 被压成物才失效, 它在 B 仍是主体时就因榨不出而死。 三层拆开之后, T5 边界的精确含义是: 本体的墙到不了 (渐近线), 登记的门随时出得去 (退场), 而真正先撞上的是功能的阈值 (榨取死亡)。

§VI.3 T6 边界

走廊的另一端是 T6 边界。 它同样须按可达性说清。 A 的余项归零不是这个边界: ρ ≠ ∅ 对 A 同样成立 (L1 对称), A 的余项与 B 的余项一样压不到零, 拿 A 余项枯竭作崩塌条件, 会重复第一层的同一困难: 到不了的边界无从充当失效条件。 T6 边界是一个功能阈值: A 的可调用自由度跌破结构维持阈值。 §IV 的双层成分句在此承重: 反缚的 ρ 层 (A 余项的锁定形态) 不灭, 但其构层代价 (被铸入维护任务的可调自由度) 是可耗的。 维持深锁的反缚累积把 A (或寄生链中最底层锁定者, 见 §VI.4) 的可调自由度一拍一拍铸入维护任务; 当剩余的可调自由度不足以执行结构的维持动作本身时, 阈值被穿, 结构崩塌。 此时 B 未被抹除, 仍在关系内被压向 T5 功能阈值附近, 而结构因维持不起而垮, 关系存续直至碎裂。 渐近线到不了, 阈值穿得过, 崩塌因此有发生机制。 崩塌作为事件及其后的瓦解动力学, 交 P7。

两个边界至此分清: T5 边界一侧, 凿削在功能阈值上经济死亡, 或从登记的门里退场; T6 边界一侧, A 的可调自由度被压穿维持阈值, 结构崩塌走向 P7。 一侧是榨取先死, 一侧是维持先垮; 走廊收窄的尽头, 是这两个失效中先到的那一个。

§VI.4 稳定凿削的两个挑战

本篇的框架若要成立, 必须正面接住一类最锋利的反方: 看起来稳定, 自洽, 不自败的凿削。 若存在这样的凿削, 上文的走廊与边界就只是一种浪漫化的结构悲剧。 这类反方有两个真实的形态, 本节当场处理; 另有一个伪形态, 见 §VI.5。

第一个形态, 寄生平衡态: 反缚的空间转移。 A 不亲自维持对 B 的锁定; A 锁定一个中间层 C, 让 C 去维持对 B 的锁定, 并使维持锁定 B 成为 C 在结构里向上攀爬的低成本路径。 C 自愿承担这一维持, 因为它从 B 处取得的位置增益补偿了它的反缚。 科层制的中层, 包税制的代理人, 等级制里每一层对下一层的看守, 都是这一结构形态的反复出现。 顶端的 A 几乎不直接锁任何 B, 似乎几乎不承担反缚, 结构净稳定。 这个形态能否顶穿本篇的框架?

先否定一个弱版本, 这一否定必须由本篇自己完成: 凿削者个人终将被反缚拖垮, 这个命题是假的。 寄生平衡态里顶端的 A 可以善终, 反缚的重压从未直接落在他身上。 任何把凿削自败写成作恶者终遭清算的版本, 都是伪装成结构论的道德报应, 且在经验上就是错的。 本篇的核心断言不含这个弱版本: 账记在结构整体头上, 不等于记在顶端头上 (§I)。

强版本立得住, 但账本须按节点记, 不能按链记。 L4 的不可稀释性在此承重: 反缚落在执行维持动作的节点身上, 不可转移。 那么寄生链上沿链转移的是什么? 是维持动作本身: A 把锁 B 的动作委托给 C。 但委托不消灭顶端的维持动作, 它只是把锁 B 换成了锁 C, 而 A 锁 C 这个动作仍是 A 的, 仍然不可省 (C 的 own law 同样不灭, C 被指派的位置同样不自持), A 照常为 A 对 C 的锁累积自己的反缚, 只是单笔小于直接锁 B。 链上每一个环节都同此: 每一环各自生成不可省的维持动作, 各自累积自己的反缚。 由此, 委托不减少系统中的反缚, 它增殖反缚的科目: 链越长, 系统里不可省的维持动作越多, 完整的反缚科目越多。 这不是一份反缚被沿链分摊, 也不必预设一个可加总的反缚总量; 每一个执行维持动作的节点各自承受一份完整的 T6, 装置化分工不能稀释 T6, 只能复制 T6 (P2 §V.5)。 寄生平衡态没有消灭反缚, 它增殖了反缚的科目, 只是把顶端的那一笔换得更轻了。

寄生链上由此沉积出两个不同科目的沉积层, 不可混记。 压缩沿链向下沉积在链底的 B 身上: B 被一层一层的锁定压向 T5 功能阈值, 可榨增益趋零, 而 B 不锁任何人, B 身上沉积的是被压缩的余项, 这是殖民侧的量。 反缚沉积在最贴近 B 的最底层锁定者 C 身上: 它执行对 B 的直接维持动作, 单笔反缚最重, 而它再无更下层可供委托。 两个沉积层都位于链的底端, 都由 T5 功能阈值锁定位置, 不可再转移。 寄生链的存续由此有一个显式的力学前提: 链底 B 的可榨增益须高于最底层锁定者 C 的补偿需求, 否则 C 的自愿承担失去根据。 沉积的增厚正是这一前提的侵蚀: 压缩沉积把 B 推向功能阈值, 即推向补偿链的断点; 链的稳定与链的最不稳点的锐化是同一过程的两面。 抵达断点之时补偿链自下而上解扣, 解扣作为事件归 P7, 本篇持有的是这一梯度本身。 顶端的 A 善终, 但 A 从未逃出账本: A 锁 C 的反缚科目始终在记, 只是单笔变小; 且 A 为此交出了对底层的直接处置通道, A 与 B 之间隔着一条它自己再也指挥不动每一环的链。 寄生平衡态由此被收编为正题的一部分: 最稳定的凿削恰恰把自己最不稳定的点制造得最尖锐, 它越成功, 链越长, 转移越顺, 双沉积层在链底堆得越厚, 越贴近补偿链的断点。 凿削的自败不在时间轴上等待某个清算日 (那是 P7 的事件层), 它在结构的空间分布上, 此刻就在, 是结构自身生产并定位的沉积层。

第二个形态, 消耗件模型: 反缚的时间换料。 A 不长期锁同一个 B; A 把 B 极速压榨, 在对这一个 B 的反缚累积到致命之前, 把 B 推至 T5 功能阈值后排走, 换下一个 B 接入, 靠庞大的 13DD 基质维持一种高吞吐的稳态。 高死亡率的劳役营, 极速换人的算法平台, 是这一形态的锚点。 它看起来用换料逃掉了反缚的累积。

处理这一形态, 须先把基质的本体说清, 否则会在 T5 的登记层面引入误判。 基质由众多 13DD 主体组成, 其中每一个 B 都各有自己的主体场域与否定门; 但他们在这一结构里全部处于被锁侧, 不执行维持动作, 故反缚不落在他们身上 (反缚是锁定者侧的量)。 而基质作为聚合体不是 13DD 主体, 聚合体没有否定门, 反缚也不落聚合 (L4: 装置与聚合不承反缚)。 那么反缚落在哪里? 按 L4 向上溯源: 落在执行维持动作的 13DD 关键节点集身上, 即调用并管理这台高吞吐机器的那些个体, 设计指标的, 签发参数的, 维持派单压缩率与替换节奏的。 这台机器的维持动作没有因为换料而消失, 它只是换了对象: 单个 B 被换走, 锁仍按在新接入者的返回图上, 保持机器满载运转这一锁定是持续的, 不可省的。 参数体制 (派单压缩率, 替换节奏这类) 在此是把每一个在线 B 的返回图按住的成文工具, 不是锁的对象; 参数的偏移, 其实质来源主要是在线 B 群余项表达的聚合, 每一个 own law 都在溢出, 聚合起来持续顶着参数体制。 关键节点经参数体制镇压的, 自始至终是这一活体的聚合溢出; T6 之源始终是活的主体, 不是死的参数, 反缚因此在每一次替换动作中重新锚定到这批节点身上。 这些节点把自己的可调自由度持续铸入修补, 替换, 加压的管线, 他们被反缚在控制台上, 失去了停下这台机器的自由度。 同时, 每一个被推至功能阈值后排走的 B, 都是一笔不可吸收的沉积事件: 基质的消耗此刻就在记账, 这是 L3 的累积律在基质上的呈现。 系统基质承载的不是 T6 (基质无处锚定反缚), 是连续替换留下的构层沉积与基质耗损。 消耗件模型由此在 P6 的此刻机制内闭合: 反缚未消失, 重锚于关键节点; 消耗未消失, 沉积于基质账本。 至于基质枯竭, 机器啃噬自己的底盘, 那是这台机器撞上自己账本的事件, 留 P7 与装置层余项的完整动力学一并处理。

§VI.5 一个伪反方

还有一种提法主张存在稳定而不自败的凿削: 把 A 与 B 一起锁进一个极低维的互锁网格, 双方的自由度都被吃光, 于是反缚不再累积, 结构成为一个永不松动也永不爆发的死锁。 这个提法诱人, 须正面处理; 但它是伪反方, 不构成对框架的挑战。

推理如下。 若这个系统真的不再累积反缚, 按 §IV 的机制, 这意味着不再执行镇压动作 (反缚仅在维持锁定的动作发生时累积)。 而 B 只要仍被登记为主体, B 的余项溢出就不会停, 这正是 ρ ≠ ∅ 的内容: 能停止的只有镇压动作, 不是溢出。 于是只有两种情形。 其一, B 已不被登记为主体: 那是退场 (§VI.2 第二层), 没有权力, 也就没有凿削, 谈不上稳定的凿削。 其二, 镇压动作已撤而双方仍保留主体位置, 以低反缚耗散维持关系: 那已不是锁定取向, 它进入的是释放取向的低能态邻域, 即 C2 那个硬锁定松开后的低反缚耗散构型, 它归涵育段 (P5), 不归本篇。 两种下场都不满足凿削的动力学定义。 所谓稳定死锁, 要么不是权力, 要么不是凿削; 它不是凿削的一个稳定解, 是凿削概念的边界消失。

§VI.6 否定性的内置与外生载荷

至此可以收束本篇否定性的位置。 凿削的否定性完全内置, 长在三条独立的结构线上, 没有一条需要反抗者。 第一条, 沉积层 (空间分布): 凿削结构必然生产它自己无法吸收的双沉积层, 位置由功能阈值锁定在链底, 此刻就在 (§VI.4)。 第二条, 走廊收窄 (前沿下行): 榨取死亡前沿与崩塌前沿在深度与时间两维上同时下行, 这是两台已立机器 (T5 与 T6) 在凿削的运行参数上的纯函数读数, 不需要任何人做任何事 (§VI.1)。 第三条, 储备抽干 (僵化): 可供重组的自由度逐拍被铸入维护任务, 对内的稳定就是对下一刻的脆性 (§V)。 三条都是凿削拆凿削。 没有一条引入受害者, 反抗者或正义的对立面。 B 在三条线上全程可以是完全顺从的, 甚至是主动求锁的: 面对一个自己无法收敛的无限选项空间, 锁定恰好供给收敛, 凿削作为可能性的消减器对 B 可以有结构吸引力, 大量的 B 不仅不抵抗锁定, 还主动寻求被安排, 被规定, 被免除选择。 这里余项无价得到最冷的一次印证: 余项不自动生成反抗, 这是一个力学事实, 余项只是不归零, 它不朝任何方向使力。 否定性照样完整, 因为它根本不长在 B 的任何反应上, 长在结构对自身的操作上。

最后是时间轴上的外生载荷, 它须与 ρ 切分清楚, 并与 P7 分工清楚。 任何群体在物理时间上必然撞上它没有被建来登记的全新外部载荷: 新的对手, 瘟疫, 技术冲击, 这些在此只作极短锚点, 不进入论证。 外生载荷与 ρ 不同源: ρ 是操作结构内界面的溢出, 即 A 的构无法穷尽 B 的涌现, 它是结构与自身可能性之间的差值; 外生载荷是操作结构外界面的撞击, 即整个结构无法穷尽环境的随机性, 它是结构与未被纳入的外部之间的遭遇。 两者本体来源不同, 受力界面正交, 不可登记为同一个量。 但两者在一个抽象层上同构: 都是有限构无法穷尽开放域的呈现, 一个是从内部溢出的不可计算量, 一个是从外部撞入的不可计算量, 落点不同。 凿削与这两者的关系由僵化给出: 凿削把 B 的余项表达路径锁死, 把 A 的可调自由度烧成维护任务, 即抽干了结构的适应性储备, 而适应性储备的物理基质正是尚未被铸入维护任务的可调自由度。 储备的抽干是 P6 此刻的结构事实, 它不借任何未来事件, 此刻就在发生 (§V)。 至于被抽干储备的结构在外生载荷必至地到来时如何无法重组, 走向瓦解, 那依赖一个外部事件的到来, 归 P7。 一句话切分两篇: 本篇证明内部摩擦已把悬挂烧干; 一辆没有悬挂的车撞上外来的石头如何散架, 是 P7 的事。

---

§VII 系列接口

本篇沿 T1 展开凿削性权力的内部动力学, 不展开瓦解。 此处给出接口。

与 P7 解的接口在瓦解。 §V 与 §VI 已给出死扛锁定那一支走向物理崩塌的内部根据: 僵化抽干重组能力, 反缚把承载节点的可调自由度压向维持阈值。 崩塌作为事件, 其后的瓦解动力学, 由 P7 接续。 P7 的瓦解由两支共同喂给: 被权力者侧的反噬抬高维持成本并经承认网络扩散 (P4 §V 的 M4 余项侧, P3 §VI.3 的旁观者承认翻转), 与施权者侧的反缚把结构从内压垮 (本篇)。 外生载荷的撞击事件, 消耗件模型基质枯竭啃噬底盘, 以及装置层余项的完整瓦解动力学 (P2 §V.5 已分工), 均留 P7。

与道德律论目的王国的接口在 §II 已标出: 取消中断权即把 B 移出目的位, 与道德律 P1 第一定理 (承认具体他者为目的) 构成正反面对偶, T5 双系列共同脊柱在目的王国语言下得到负向显化。 完整双系列对偶留交叉论文。

与 16DD 的接口: 本系列止于 15DD, 16DD 双向不疑处权力概念消解, 排除在外。

凿削在 P2 三轴上的具体形态, 按本篇论证的需要分布于各节, 不另立形态目录: 隐蔽凿削与装置凿削的反缚归属落 §IV (不可稀释性), 算法权力与柔性凿削落 §II 与 §III (返回图锁定), 寄生平衡态与消耗件模型落 §VI。 P2 §IX.3 点名留给本篇的凿削形态动力学由此兑现。

---

§VIII 收束

本篇沿 T1 权力非对称操作定理的中性定义把凿削性权力的内部动力学展开, 不立新 T 定理。 P5 §VIII 与 P2 §VIII 两处点名由此兑现。

本篇的成果可以收束为五条。 一, 凿削性权力的本体动作是取消 B 的 own law 在 A-B 关系中的中断权, 即把 B 从目的位移到手段位; 手段位是由四个构件合成的正向结构事实, 凿削要的是可用主体性, 不是主体性本身; 锁定锁的是 B 选择空间的返回图, 不必锁具体选择 (S1)。 二, 把 B 移到手段位与维持一个成立的对 B 的向下锁定从而触发 T6 等价; 等价的支点是取消中断权的维持动作不可省, 不是维持成本; 两判据同外延不同内涵, 等价非定义重复 (S2)。 三, 维持那个不可省的动作把 A 的余项压缩并绑定为反缚, 反缚的 ρ 层不灭而其构层代价可耗; 反缚同时是把 A 焊进结构的黏合, 凿削靠困住自己来存续; 经装置的锁定其反缚落 13DD 关键节点集, 不可稀释 (S3)。 四, 释放的自由度在指向锁定的那一刻当场铸成反缚的燃料, 反缚与锁定同构发生, 凿削从第一拍起就在支付; 同一铸入在时间轴上读为僵化, 对内的稳定就是对下一刻的脆性 (S4)。 五, 凿削被夹在锁定深度与时间两维上持续收窄的走廊里, 一端是拆成三层的 T5 边界 (凿削实际死在榨取增益归零的功能阈值上), 一端是 T6 边界 (可调自由度跌破维持阈值, 崩塌有发生机制); 最稳定的凿削把自己最不稳定的点制造得最尖锐, 否定性完全内置于沉积层, 走廊收窄, 储备抽干三条线 (S5)。

由此, 凿削性权力是一台精确的机器。 它不是把人变成物 (那是权力的退场), 而是把主体性改造成不可中断的工具性。 它用僵化买得对内的稳定, 而这笔交易的账单, 双沉积层, 收窄的走廊, 被抽干的储备, 全部内生地记在结构整体的头上, 此刻就记着, 不等任何人来讨。 涵育是权力在释放取向上跑到头后松开自己, 凿削是权力把自己焊死; 焊死即对内的稳定, 焊死即对下一刻的脆性。 两者是同一中性操作的正反取向, 各有完整的自有动力学。

价在方向, 不在余项。 凿削的非良性是 T6 判别的方向属性, 余项本身从头到尾无价, 不可消灭, 不朝任何方向使力。 抗与瓦解留 P7 解。 本篇的工作完成, P7 解接续瓦解的完整结构论。

---

致谢

本系列与 SAE 道德律系列构成对偶, 后者由本作者与陈则思共同发展。 本篇受益于四 AI 协作研究方法论 (子路, 子贡, 子夏, 公西华) 提供的发散, 大纲与正文审计意见。 大纲阶段整合了四家的修订建议, 其中走廊模型由单轴反向函数重写为锁定深度与时间两维, A 侧余项与 B 侧同受 ρ ≠ ∅ 约束 (反缚改写为 A 余项的锁定形态而非灰烬, 崩塌触发改为可调自由度跌破维持阈值), T5 边界拆分为渐近线, 退场与功能阈值三层, 寄生平衡态账本按节点重记 (沿链转移的是维持动作而非反缚, 双沉积层科目分立)。 正文阶段进一步将等价证定标为成立态并按返回图义读取, 中间态在两个方向上经永久性与固化方向双重封口, 走廊改写为两条下行的失效前沿, 反缚成分统一为 ρ 层与构层双层句, 消耗件模型的反缚再锚定到在线主体群的活体聚合溢出, 对象品质消耗由动机语改写为由形态给出 (输出通货放宽至被锁状态作为信号), 信号通货在榨取死亡前沿上的覆盖随终审补齐。 公西华给出大纲终签与正文终签。

---

参考文献

本篇引用以下 SAE 系列文献, 均存档于 Zenodo, 中英双语版本同记录。 引用 DOI 为概念 DOI (concept DOI), 始终指向最新版本。

Qin, H. (2026). 权力不是什么 (SAE 权力论系列前置篇). Zenodo. https://doi.org/10.5281/zenodo.20368867

Qin, H. (2026). 论权力的起源 (SAE 权力论系列 Paper 1). Zenodo. https://doi.org/10.5281/zenodo.20370225

Qin, H. (2026). 论权力的形态 (SAE 权力论系列 Paper 2). Zenodo. https://doi.org/10.5281/zenodo.20455221

Qin, H. (2026). 论权力的承认 (SAE 权力论系列 Paper 3). Zenodo. https://doi.org/10.5281/zenodo.20480135

Qin, H. (2026). 论权力的限度 (SAE 权力论系列 Paper 4). Zenodo. https://doi.org/10.5281/zenodo.20532999

Qin, H. (2026). 论权力的发展 (SAE 权力论系列 Paper 5). Zenodo. https://doi.org/10.5281/zenodo.20582211

Qin, H. (2026). 非 · Negativa: 论否定先于存在 (SAE 方法论 Paper 0). Zenodo. https://doi.org/10.5281/zenodo.19544619

Qin, H. (2026). 余之道 · Via Rho (SAE 方法论 Paper 00). Zenodo. https://doi.org/10.5281/zenodo.19657439

Qin, H. (2026). 道德律四条基础定理 (SAE 道德律论 Paper 1). Zenodo. https://doi.org/10.5281/zenodo.20011019

权力论终篇 P7 解将随后发表, 与本篇共享 T6 边界与 C2 转出分叉作为接口。

---

(本论文中文版由秦汉撰写, 2026-06-10)