【(Preliminary)Perspective】A First Discussion of Power and Value in International Relations from a Generative Relational Perspective

ENGLISH

A First Discussion of Power and Value in International Relations

from a Generative Relational Perspective

Wanhong Huang (huangwanhong@serendip.ngo)

Abstract

Theories of international politics measure power, aggregate it, and ask what follows from its distribution. They seldom ask why there is any. Power admits of derivation. A process that generates must differentiate, since a generation that leaves the world indistinguishable from what it was has generated nothing; what it differentiates includes the capacity to supply the terms on which coordination proceeds; and a relation whose terms are supplied by one of its parties is a relation with a direction. Asymmetry is therefore a condition of relations in which something is made, rather than a corruption of them, and it resists removal that leaves generation standing. The question of justice cannot then be how to eliminate asymmetry, and becomes instead what the asymmetry does to the relation in which it stands.

Answering that question in international politics runs into an obstacle that the available frameworks do not meet. A state supplies technology and receives recognition, supplies a guarantee and receives basing rights. Between such goods there is no ratio, and for the account to close, something must be made to count as repayment for something of another kind. That operation requires a rate, and the rate is not given by the values it settles. It is posited, and there is no correct one. Three evasions are closed. Monetisation performs the conversion it claims to avoid; categorial separation forbids the question it claims to answer; subjective satisfaction consults the effect of the rate as the test of the rate.

That the setting of a common metric among unlike things is a mode of power belongs to the sociology of commensuration and is inherited here rather than claimed. But that literature theorises the boundary: a metric advances upon a sphere held apart from it, and resistance takes the form of refusing the metric outright. Refusal presupposes a position from which one may withdraw and remain what one was. A state that cannot withdraw without ceasing to count as a party has no such position, and is settled against notwithstanding its objection. Where refusal is foreclosed, commensuration ceases to be a contested boundary and becomes a constitutive condition of the relation, so that the international case is not one further instance of the sociological phenomenon but its limit, at which the phenomenon changes character.

Two consequences follow. The colony whose material ledger runs positive, a case on which generative justice, unequal exchange, and the sociology of commensuration are alike silent, is diagnosed: the injustice lies neither in the flow nor in the unfairness of the rate but in the foreclosure of any position from which the rate might be reopened. And the normative condition of a settlement is located in the contestability of its rate, which is not a proceduralist retreat but the residue of resistance once refusal has been foreclosed. The argument surrenders the classical vocabulary of exploitation, which requires a common measure that the goods of international politics do not admit. The paper closes on the question it does not answer, which is who reorganises a loop from which no party may withdraw.

Keywords: power; asymmetry; commensuration; generative justice; balance of power.

AI usage statement: The ideas, claims, and theoretical framework of this paper originate with the author. AI tools were used as an aid in drafting, in the location and verification of empirical and bibliographic sources, and in revision. The origin of each idea has been verified by the author, who takes full responsibility for the content.

§1 Introduction

Balance-of-power theory treats power as a stock. It measures it, aggregates it across the units of a system, and asks what follows from the way it is distributed. It asks whether a distribution is stable, whether a concentration is dangerous, whether a transition tends toward war. It seldom asks why there is any power to distribute.

The omission has consequences beyond the point at which it occurs. A theory that receives power as a primitive is poorly placed to say what kind of thing power is, since it has no account of what power is a kind of, and a theory that is silent on the kind is confined to the register of quantity. It can say that there is more here and less there, that the more is threatening and the less is vulnerable, and that a prudent policy inclines toward the equalisation of the two. What it struggles to say is whether a given exercise of power built something or consumed something, for that is a question about the character of a relation and the theory has the magnitudes alone.

Power admits of derivation. A process that generates must differentiate, since a generation that leaves the world indistinguishable from what it was has generated nothing. What it differentiates includes the capacity to supply the terms on which coordination proceeds, for the supplying of terms is itself a productive act and falls under the same differentiation as everything else produced. And a relation whose terms are supplied by one of its parties is a relation with a direction. Power is what that direction is called.

Asymmetry appears on this account as a condition of a relation in which something is made, rather than as a corruption of relation, and it resists removal that leaves generation standing. Two consequences follow at once. The tragedy on which realism has long insisted requires neither an appetite for domination nor a condition of anarchy: it would arise in any system that produced anything. And the programme that proposes to abolish asymmetry is self-defeating in a manner that is structural rather than practical, since in the degree to which it removed difference it would remove the differentiation in which generation consists.

The question of justice cannot then be how to eliminate asymmetry, and becomes instead what the asymmetry does to the relation in which it stands. Some asymmetries raise the capacity of both parties to generate independently; others raise the output of the weaker party while lowering its capacity to produce on its own account, which is how dependency is ordinarily manufactured, through provision rather than through deprivation. The distinction is invisible to a theory that reads magnitudes, and it is the distinction on which everything else in this paper turns.

Asking whether value returns to those who generate it therefore becomes the right question to ask, and in international politics it runs immediately into an obstacle. A state supplies technology and receives recognition. It supplies capital and receives legitimacy. It supplies a security guarantee and receives basing rights. It supplies an apology and receives, or fails to receive, the closure of a grievance. What goes out and what comes back belong to different registers, and between the registers there is no common substance and no ratio. For such a loop to close, something must be made to count as repayment for something of another kind. That operation requires terms, and the terms are not supplied by the values themselves, since values of different registers stand in no relation of magnitude to one another. The terms are therefore posited, and once it is seen that they are posited, it must be asked who posits them.

The obstacle is general and it defeats the available frameworks severally. Realism does not meet it, having converted the whole of the difficulty into a single scalar before the difficulty can arise. The theory of unequal exchange meets it and answers it with a common measure, labour time, which is unavailable for guarantees, recognition, and apologies. And the theory of generative justice, which has established that value ought to return to those who generate it, was built in domains where return is homogeneous: the value produced by a craft returns to the craftsman, as income or skill or ownership or time, and one may inspect whether it did. Where what returns is of a kind with what was given, the account is kept in one currency, and no rate of conversion is required. Carried onto terrain where the registers differ, the requirement that value return becomes a requirement that cannot be checked until someone has said what counts.

The question of who sets the terms on which unlike things are made to settle against one another is an old one, and much of the ground has been taken. Espeland and Stevens established, in 1998, that the setting of a common metric among unlike things is a mode of power, that the power is constitutive rather than merely strategic, that it grows invisible through routinisation, and that political bargaining presupposes it. Postcolonial criticism established that what returns to the colonised returns in the coloniser’s language. Constructivism established that meaning precedes interest and that institutions are sedimented practice. The relational theories of world politics share with the present argument the premise that relations precede entities. And Eglash established that value ought to return to those who generate it, which is the claim this paper carries onto terrain where it cannot be checked. The debts are large, they are acknowledged throughout, and the argument that follows is what remains once they are paid.

The sociology of commensuration is a theory of the boundary. Its cases are cases of encroachment, in which a metric advances upon a sphere that has been held apart from it, and the political drama is the drama of whether the advance shall be permitted. Resistance, in that literature, tends to take one characteristic form, and it is the claim of incommensurability: the assertion that this is not the kind of thing that may be traded against that. The Yavapai make it of their ancestral land. Homemakers make it of their lives. Movements make it of their commitments when they decline the bargaining table. It is the strong form of refusal, and where it can be sustained it is the apt one, since it does not dispute the terms of an accounting but declines to have the accounting at all. But the strong form requires a position from which it may be held. The party must be able to walk away and remain what it was.

A state is rarely placed to do this. It may declare its sovereignty incommensurable, and it will be settled against notwithstanding. It may hold that no quantity of investment discharges the obligation incurred by a colonisation, and the ledger will be drawn up regardless, the investment entered on the one side and the extraction on the other, and the account declared square by parties who did not require its assent. The declaration is heard. The account is closed anyway. The reason is not that states are weaker than communities, which they are not. It is that to refuse settlement is to withdraw from the field within which settlements are made, and to withdraw from that field is to become, not a state that has declined to be settled against, but a state that does not register as a party.

Where refusal is foreclosed, commensuration ceases to be a contested boundary and becomes a constitutive condition of the relation. The international case is then not one further instance of the phenomenon that sociology has described, but its limit, at which the phenomenon changes character.

The argument costs something that is better declared at the outset than discovered in the course of things. If the rate of settlement is posited rather than discovered, then there is no correct rate, and a critique that condemns a settlement for having been struck at the wrong rate has helped itself to a measure it does not possess. This removes the classical vocabulary of exploitation, which works by comparison to a true value and requires that a true value be available. The theory of unequal exchange possesses such a measure in labour time. No comparable measure is available for the goods that constitute international politics, which are not produced by labour and are not quantities of any substance, and other ground must therefore be found.

The ground proposed is contestability. Where a true rate is unavailable, the one property a rate can have that is not itself a posit is its openness to being reopened by those it binds. The features of a rate’s content were put there by somebody. That it may be questioned was not. And since the strong form of resistance, the refusal of the metric altogether, is foreclosed to a party that cannot withdraw, contestation of the terms is not a weaker substitute chosen for convenience. It is what remains.

One consequence concerns the discipline within which the argument is offered. Realism converts recognition, justice, identity, culture, and legitimacy into a single scalar, and holds that these bear upon the conduct of states in the measure, and largely only in the measure, that they can be so converted. That is itself a settlement across registers, performed at a rate, and realism’s claim to be a science of international politics rests upon its not acknowledging as much. It presents the conversion as a description of what moves states, where it is better understood as a stipulation about the terms of the account. That everything else can be settled against capability is less an observation about the world than the founding posit of the field.

The paper proceeds as follows. The literature is classified, not by school, but by what each tradition does with asymmetry, and the classification locates a gap while conceding what fills the neighbouring ones. Power is then derived from generation, and the derivation is turned against realism, against egalitarianism, and against the relational theories that share the premise of this argument and draw from it the conclusion that relations are fundamentally harmonious. A criterion is supplied for distinguishing the asymmetry that builds from the asymmetry that consumes. The heterogeneity of international value is established, and with it the unavoidability of a posited rate, by closing the three evasions that might otherwise avoid it. The consequences of foreclosed refusal are drawn out and applied to the colony whose material ledger runs positive, a case on which the rival frameworks are silent. A mechanism is proposed, at a lower level of confidence, for why the weak accept rates they can perfectly well see to be arbitrary. A normative condition is offered in place of the one surrendered. A formal representation is sketched and its difficulties stated. The paper closes on the question it does not answer, which is who is to reorganise a loop from which no party can withdraw.

§2 The Literature on Asymmetry

This section does not survey schools. It classifies theories by what they do with asymmetry, and the classification is itself an argument. Three questions organise it.

Q1. The genetic question. Whence asymmetry?

Q2. The normative question. Does asymmetry as such constitute injustice?

Q3. The settlement question. When the values exchanged belong to different registers, how is the account settled?

The thesis of the section is that Q3 has been answered principally in one place, the sociology of commensuration, and answered there for the case in which the parties may refuse the metric. The case in which refusal is unavailable, and settlement therefore compulsory, has received little attention. That case is the ordinary case of international politics.

Asymmetry as Given

The realist family takes asymmetry as the starting point of analysis rather than as a thing to be explained. Classical realism grounds it in the appetite for domination; structural realism in the distribution of capabilities; offensive realism in anarchy compounded by uncertainty about intentions; the balance of threat in perceived threat rather than raw capability; power transition theory in differential rates of growth. Hegemonic stability theory, alone among them, permits the concentration of capability to be beneficial.

The difficulty at Q1 runs deep and is not a matter of emphasis. Waltz’s distribution of capabilities redescribes the explanandum rather than explaining it. Power transition theory’s appeal to differential growth restates the explanandum: asymmetry arises because things grow at different rates, and growing at different rates is what asymmetry consists in. Morgenthau’s appeal to human nature is the tradition’s one sustained attempt at a genesis, and it is an anthropological hypothesis rather than a structural one, which leaves it poorly placed to explain the emergence of power in a system whose members are states rather than persons, except by an analogy that is seldom defended.

There is besides an internal difficulty in structural realism which this paper takes seriously and which the tradition has not resolved. Waltz declares structure to be constituted by the positional relations among units, and then computes the third of its three elements as the sum of unit attributes. Structure is claimed to be relational and is calculated as substantialist. This paper takes the relational claim seriously and follows it to consequences other than those Waltz drew.

Hegemonic stability theory deserves separate treatment, since it lies near to the present argument and is a friend rather than an adversary. It concedes what the rest of the family does not, that a concentration of capability may strengthen the system in which it is concentrated, when the concentrated party supplies public goods rather than extracting rents. But it offers no criterion by which the two cases may be told apart. It distinguishes them after the event, and by their outcomes, which is a record rather than a criterion. The criterion advanced in the following section is offered to hegemonic stability theory as the thing it has been missing.

Asymmetry as Governable

The liberal family treats asymmetry as a hazard to be neutralised. Institutions reduce transaction costs and supply information; regimes stabilise expectations; democratic peace locates the danger in regime type rather than in the differential itself.

Keohane and Nye must be taken seriously, since Power and Interdependence lies close to this paper’s concerns. Their insight is that asymmetry in a relation of interdependence is a resource of power within it, and not merely a condition of it, and that this power is detachable from military capability. The insight is correct and it is important.

Its limitation at Q3 is nonetheless exact. Asymmetric interdependence is theorised as bargaining leverage, and bargaining leverage is a matter of which party better tolerates the severance of the relation. This is a scalar, a measurement along the single dimension of tolerance, and it presupposes that the matters in dispute have been rendered comparable enough to be bargained over. It answers the question of who holds leverage. It does not pose the question this paper poses: when one party supplies a security guarantee and the other supplies basing rights, who owes whom, and by what measure? The framework converts a heterogeneous relation onto a single dimension without remarking that it has performed a conversion.

The difficulty at Q2 belongs to the family as a whole. Asymmetry is assumed, with little argument, to be something for institutions to neutralise. That it might be the condition of generation rather than a threat to it is seldom considered, and where the institutional aim is levelling, the institutional achievement tends toward the suppression of what the levelling was to have distributed.

Asymmetry as Structural Domination

Unequal Exchange

The theory of unequal exchange lies close to this paper’s central question, and the relation between them must be stated with precision if the paper is not to be dismissed as a rediscovery.

Emmanuel’s mechanism is the differential of wages. With capital mobile and labour immobile, and with wages institutionally determined and therefore exogenous, the periphery sells its products below their social value and the core sells above, and surplus is transferred in the direction of the core. Amin extends the account. What Emmanuel condemns is a rate, and the rate he condemns is denominated in labour time.

The first difference is that Emmanuel possesses a common measure and this paper does not. Labour time is the invariant substance against which the price is found wanting, and it is what permits the exchange to be called unequal. Without a measure the concept is difficult to form, since inequality is a relation to a metric and there is little to be unequal in where no metric obtains. The difficulty this paper faces is therefore harder than Emmanuel’s, and it is not incidental: between technology and recognition, between a security guarantee and a basing right, between capital and legitimacy, there is no common substance, and not merely none yet discovered.

The second difference follows. Where a true value is available, one may say that the correct rate is one thing and the actual rate another, and that the difference is exploitation. The critique targets the deviation of the rate from its truth. This paper denies that a true rate is available, and is therefore unable to appeal to a correct one. Its normative fulcrum must be placed elsewhere, and the placing of it is the subject of a later section.

The surrender of the common measure is a vulnerability, and it is stated as such. To abandon the measure is to abandon the classical definition of exploitation. The cost is real and it is paid deliberately.

The reason for paying it is that the labour theory of value is ill suited to the terrain of international politics. Security guarantees, recognition, legitimacy, strategic position, and historical apology are not produced by labour time and are not quantities of a substance. A critical theory that depends upon a common measure falls silent where international politics is thickest. Emmanuel can tell us a great deal about coffee and textiles. He can tell us little about whether the Marshall Plan was repaid, about what would repay a colonial apology, or about whether a state has purchased its recognition too dearly. These are not the marginal cases of the field. They are among its central ones.

It should be recorded that this paper is not alone in its abandonment. Hornborg has objected, against the recent revival of ecologically unequal exchange, that the labour theory of value which guided Emmanuel and Amin can be conceptualised only in monetary terms, and that to infer a transfer of value from a discrepancy in wages is to reify value into a substance created at particular points in the system. He asks how value could be represented in any metric other than money, and treats the absence of an answer as a difficulty for the theory. This paper takes the absence as its datum. Where Hornborg finds the want of a metric a problem for critique, this paper finds it the starting point of one.

There is a further asymmetry between the two frameworks which will matter below. Emmanuel requires a negative material balance in order to detect exploitation, and has therefore little to say about the case in which capital flows inward.

Gramscian Political Economy

Cox’s hegemony is coercion together with consent, materialised in a historic bloc and reproduced through ideas, institutions, and material capabilities. It comes close to a theory of symbolic power over structure, and it is a considerable achievement.

The difference is this. Cox theorises how a particular order comes to be accepted as legitimate. He does not theorise the operation by which a good of one kind comes to count as settlement for a good of another. Legitimation is the consequence of that operation; Cox describes the consequence with great insight and leaves the operation unsupplied. He can tell us that the postwar order was consented to. He tells us less about how the provision of a security order came to be an acceptable answer to the extraction of monetary autonomy, which is to say, how a good in one register came to be received as payment for a good in another.

Postcolonial Theory

Postcolonial criticism has established much of the substance that this paper’s central sections describe. Spivak’s epistemic violence, Chakrabarty’s provincialisation of Europe, and Fanon’s account of the colonised subject converge upon a single observation: what returns to the colonised returns in the coloniser’s language, and the language in which the return may be assessed is the language of the party that made it.

The discovery belongs to postcolonial criticism. What is claimed here is narrower, and must be kept narrow.

The insight has been developed as cultural and epistemological critique. What is attempted here is its reconstruction as a formal mechanism within a system of value circulation, specifically as the unilateral positing of the rate at which a good in one register discharges a debt incurred in another. The purpose of the reconstruction is not to improve upon the critique, which stands in little need of improvement. It is to render the critique connectable to the theory of generative justice on one side and to institutional design on the other, which cultural critique in its own idiom is not. Whether this is a gain or a loss is left to the reader. It is a gain in tractability and a loss in thickness, and both should be said.

Asymmetry as Constituted

Constructivism established that meaning precedes interest, that anarchy is what states make of it, and that institutions are the sediment of practice. The English School established that international society possesses common norms and institutions. Relational theories, in the work of Jackson and Nexon and of Qin, have established that relations precede entities. The affective and practice turns have established that emotion and background competence are constitutive of conduct rather than residual to it.

This paper contests little of this and claims none of it. The temptation to position the present framework as the theory that explains how value emerges, in contrast to a generative justice that explains only how value flows, is to be refused. That position is occupied, and it has been occupied since 1992.

Qin’s relational theory requires separate treatment, because it asserts as its ontological premise exactly what this paper’s programme asserts, that relations precede entities and that actors are actors-in-relation. Its metaphysical core is the correlated relationship of 阴 (yin) and 阳 (yang), whose basic state it posits as harmony, and from that posit it infers that all relations are fundamentally harmonious and are able to be managed as such. The 中庸 (zhongyong) dialectic supplies the epistemological schema.

The standing objection to relational theory, which Qin has addressed directly, asks how a theory beginning from relation rather than from opposed substances can identify subject and object at all, and how it can then give any account of the relation of domination and subordination between them. Qin’s answer is to dissolve the dichotomy: the appearance of irreducible opposition is an artefact of a dualistic lens, and a dialectic that assumes no binary in the first place refuses the terms in which the difficulty was posed.

The answer given in this paper is the opposite one, and it is given from within the same premise. It is set out in the following section.

The difficulty of this class at Q3 is general and structural. These are theories of the generation of meaning rather than theories of circulation. They seldom ask whose meaning counts when a cross-register account must be settled, and they offer little account of the settlement function of meaning, which does not follow from an account of its generation and is not contained in one.

The Sociology of Commensuration

The tradition that has answered Q3 lies outside international relations, and it must be given its due at length, since the greater part of what this paper might have claimed for itself belongs to it.

Espeland and Stevens define commensuration as the expression or measurement of characteristics normally represented by different units according to a common metric. It transforms qualities into quantities and difference into magnitude. Their account establishes, and this paper inherits rather than asserts, the following. Commensuration is a mode of power, and its power is both strategic and constitutive. It is not the neutral technical process its advocates claim, and it reconstructs relations of authority, creates new political entities, and establishes new interpretive frameworks. Its operation becomes invisible through routinisation, so that the labour it requires and the assumptions surrounding it are forgotten. It constitutes what it purports to measure. Marx is read within their account as holding that labour is the great commensurator and that value commensurates by abstracting away the distinctiveness of particular work, which is the reading this paper has given of Emmanuel and which is therefore not this paper’s. And politics itself, understood as the art of compromise, is on their account a broad instance of commensuration, so that interest-group bargaining presupposes the very operation whose terms it is not placed to examine.

Within international relations the concept has entered by a narrow door. Broome and Quirk employ it in their account of global benchmarking, where commensuration figures as one step in a chain of numerical translation, alongside simplification, extrapolation, reification, and symbolic judgment. But the object of that literature is the indicator and the ranking: the translation of normative values into scores by which distant governance is conducted. Its object is not the settlement of a cross-register account between two parties, and the question this paper asks does not arise there.

Two limits of the commensuration literature define the space in which this paper works, and they must be stated exactly.

The first is that it declines the normative question by design, and the design is a virtue in a review article and a limitation here. Commensuration, on its account, may enlarge decision-making or may legitimate a decision taken in advance; it may be manipulated by elites or may constrain them; it may arm dissenters or produce disciplined subjects. The even-handedness is scrupulous and deliberate. What it means is that the literature is not placed to say when a commensuration is extractive, and this paper must be.

The second limit concerns not what the literature says but the direction in which it looks.

The cases from which the sociology of commensuration is built are cases of encroachment. A metric advances upon a sphere that had been held apart from it, and the political question is whether the advance shall be permitted. Resistance accordingly tends to take one form: the claim of incommensurability. The Yavapai say of their land that it is not the kind of thing that may be traded against a dam. Espeland and Stevens even name the diplomatic instance, observing that the territorial commitments of the Israeli and Palestinian peoples have confounded countless efforts to commensurate territorial interests at diplomatic bargaining tables, and the case appears in their account as a case of refusal.

It is a theory of the boundary. Its account of resistance rests upon the possibility of saying: not this, not here, not on those terms.

International politics, in the greater part of its ordinary operation, takes place in the interior.

Asymmetry as a Governable Condition of Generation

The final class is sparsely populated within international relations and exists largely outside it. Eglash’s generative justice holds that asymmetry is not injustice and that the severing of the value loop is. Ostrom supplies the design principles of self-governance without thematising asymmetry. Sen and Nussbaum move the criterion of justice from resources to capabilities. Mauss and Sahlins establish the obligatory and relation-producing character of the gift. Polanyi establishes that not all integration is market exchange and that reciprocity and redistribution run upon different logics.

Sen’s move deserves acknowledgment, since the criterion advanced in this paper is its cousin. The difference is twofold. Sen’s capability is an attribute of individuals and a matter of functionings and freedoms, of what a person is able to be and to do. The generative capacity employed here is relational, being a property of a position within a circuit rather than of a node, and it is measured across a closed circuit rather than at a point. The question is less what a party is able to do than whether, after one complete circuit of the relation, it is more or less able to generate independently than it was.

Polanyi is a near antecedent to the recognition that value circulates in more than one register. He did not ask how the registers settle against one another, and the question does not arise within his framework, but it was he who made it possible to see that there are registers.

Eglash’s limit is the limit of his domain, and it is not a defect. His cases are the braiding shop, the cooperative, the community of makers, and in these the return is homogeneous: the value generated by labour returns to the labourer, as income, as skill, as ownership, as time. He therefore has little occasion to require a rate of conversion, and little occasion to encounter the power that posits one. This paper begins where his domain ends.

Positioning

This paper displaces little. It observes that a question difficult to avoid in international relations has been left open at the junction of two literatures that do not meet.

The theory of generative justice supplies the norm and presupposes that return is homogeneous. The sociology of commensuration supplies the account of heterogeneity, establishes that settling it is a mode of power, and declines the normative question by design. Neither reaches the case in which settlement is compulsory because refusal has been foreclosed.

And since no true rate is available, this paper is unable to appeal to a fair one. The normative fulcrum must be moved. That difficulty is the paper.

§3 The Genesis of Power

Realist theory takes power as a primitive. It measures power, aggregates it, distributes it, and asks after the consequences of its distribution, and it seldom asks why there is any. The omission is not a small one. A theory that is silent on where power comes from is poorly placed to say what kind of thing it is, and a theory that is silent on the kind is largely confined to asking how much of it there is here and how much there. This section supplies the derivation that realism lacks, and the derivation has consequences that realism is unlikely to welcome and that egalitarianism will welcome still less.

The Derivation

The chain runs from generation to power in five steps.

If a process produces nothing that was not there before, it is not a process of generation. This is a matter of what the word means, and little turns on it beyond the fixing of terms.

If what is produced introduces no difference, it has not been produced. A change that changes nothing is not a change, and a generation that leaves the system indistinguishable from what it was has generated nothing. Generation therefore entails difference, as a condition of its being generation rather than as a frequent consequence of it.

Difference among the parts of a system is what permits those parts to stand in relation. Where the parts are indistinguishable there is nothing for a relation to obtain between. Relation therefore presupposes difference, and generation supplies it.

Parties that stand in relation must coordinate, and coordination proceeds upon terms. The terms are not given with the relation. They must be supplied, and the supplying of them is a further question which the fact of the relation leaves open.

And where the terms of coordination are supplied by one of the parties rather than jointly, the difference between the parties acquires an effect that runs in a direction. This directed effect is what the word power names.

The Provenance of the Terms

The fourth and fifth steps are where the argument is most exposed, and the exposure should be met rather than passed over. It may be granted that coordination requires terms, and granted also that terms unilaterally supplied produce a directed effect, while the intervening claim is denied. Why should the terms be supplied by one party rather than settled between them? Nothing in difference as such appears to require it. Two parties may differ and may negotiate the terms of their coordination as equals, and if that is the ordinary case, then power is a contingency of bargaining rather than a consequence of generation, and the derivation collapses into a description of unlucky circumstances.

The reply is that the supplying of terms is not a further activity standing outside the generative process, to be distributed between the parties by some arrangement external to it. The supplying of terms is itself a generative act. To furnish the terms on which a relation shall proceed is to produce something that was not there before, and to produce it is to exercise the very capacity whose unequal distribution the preceding steps have established.

Generation produces difference, and what it differentiates is the capacity to generate.

The terms of coordination are themselves something generated. The party better placed to generate is therefore better placed to generate the terms.

Unilateral conditioning is accordingly not a supplementary assumption laid upon difference from outside. It is difference, applied to the coordination problem itself.

The point may be put as a dilemma for the objector. Either the terms of coordination are produced, or they are found. If they are found, we must be told where: they are not in the parties, since the parties differ and each would furnish different terms; and they are not in the relation, since the relation is what the terms are required to constitute. If they are produced, then their production falls under the account of production already given, and is subject to the differentiation that account establishes. The objector who wishes the terms to be settled jointly and as equals must therefore hold that the parties, unequal in their capacity to generate everything else, are equal in their capacity to generate this. No ground has been offered for the exception, and it is difficult to see what ground there could be.

It does not follow that the party with the greater capacity always supplies the terms, and the claim is not that it does. Capacities may be unequal in one respect and reversed in another, and a party weaker in production may be stronger in the particular production of terms, which is a matter of language, of precedent, of the possession of a forum. What follows is weaker and sufficient: that the supply of terms is subject to the same differentiation as everything else generated, and that a symmetrical supply would require a symmetry the system does not otherwise exhibit. Joint settlement between equals is therefore not the default from which unilateral conditioning departs. It is an achievement, and where it occurs it is because something has been done to produce it.

The Middle Term

One clause carries the argument, and the argument fails without it.

Difference is not power.

Difference under unilateral control of the terms of coordination is power.

The clause must be insisted upon, since without it the derivation slides toward the position that all difference is power, and a concept that applies to everything discriminates nothing. That slide is a familiar hazard of the poststructural tradition, and this paper declines to take it. Two parties may differ in many respects and stand in a relation whose terms neither supplied alone, and in such a relation the difference is fertile and is not power. What converts a difference into a power is neither its magnitude nor its persistence. It is that the terms on which the parties coordinate were furnished by one of them.

Difference is accordingly the potential of power rather than its actuality. The distinction is what permits the theory to say, later, that some asymmetries generate and others extract. A theory that identified difference with domination could not say this, since on such a theory every relation would be a domination and the word would have ceased to do work.

The reader may notice that the argument has begun to describe the operation with which the paper is principally concerned. The terms on which unlike things are made to settle against one another are terms of coordination, supplied by a party, and the account of their provenance given here is the account that will be required again when the rate of conversion is examined. The two are the same structure met at different depths, and the later sections may be read as the working out of what has here been established in general.

Two Propositions

Proposition 3.1 (Genesis of Power). Power does not originate in the possession of resources. It emerges from relational differentiation within generative processes. Generation produces difference, and what it differentiates includes the capacity to supply the terms on which coordination proceeds; the terms so supplied lend the difference a direction, and the directed difference is power. Power is therefore to be expected in a generative relational system, and its occurrence does not depend upon greed, fear, or intention on the part of any actor.

Proposition 3.2 (Ineliminability of Asymmetry). A system without difference is a system without generation. The elimination of asymmetry and the maintenance of generation are not jointly achievable.

Three Consequences

For Realism

The tragedy on which realism has long insisted is relocated and, in being relocated, deepened. It arises neither from the human appetite for domination, as the classical realists held, nor from the condition of anarchy, as the structural realists hold. It arises from generation, and it would arise in a system that produced anything, however its members were constituted and however authority above them were arranged.

The claim is at once friendlier and more radical than realism’s own. It is friendlier because it removes the imputation of vice: no appetite need be posited, and the power that arises among states arises as it would among parties of any kind who produced and exchanged. It is more radical because it forecloses the standing hope of the discipline’s other half. If power is the price of generation, then institutional arrangement is unlikely to escape it, and the world state, arriving at last, would find that it had exported the problem to its own interior rather than solved it.

For Egalitarianism

The programme that proposes to abolish asymmetry is not merely difficult. It is self-defeating in a manner that is structural rather than practical. In the degree to which it succeeds in removing difference, it removes the differentiation in which generation consists, and therefore removes the production of the value whose distribution was the point of the exercise. A perfectly level system is a system in which little is made. This does not condemn the egalitarian impulse, which this paper shares. It condemns one construal of its object, and it obliges the impulse to find another, which the following section attempts to supply.

For Relational Theory

Relational theories of world politics have advanced upon a premise this paper shares, that relations precede entities and that actors are actors-in-relation rather than substances that subsequently enter into commerce. Qin’s relational theory is the most fully developed of these, and it takes as its metaphysical core the correlated relationship of 阴 (yin) and 阳 (yang), whose basic state it holds to be harmony, concluding that relations are fundamentally harmonious and are able to be managed as such. The claim is grounded in a long tradition of reflection upon the relational character of the world, and it is not advanced lightly.

The standing objection to relational theory, which Qin has addressed directly, asks how a theory that begins from relation rather than from opposed substances can identify subject and object at all, and how it can then give an account of domination and subordination between them. Qin’s answer is to dissolve the dichotomy. The appearance of an irreducible opposition between the dominating and the dominated is, on his account, an artefact of a dualistic lens, and the 中庸 (zhongyong) dialectic, which assumes no binary in the first place, dissolves the tension by refusing the terms in which it was posed. The move is coherent and it is available.

This paper accepts the same relational premise and returns the opposite answer.

Domination need not be an artefact of the lens. It is derivable from the premise. Relations that generate must differentiate; what they differentiate includes the capacity to supply the terms on which the relation proceeds; and a relation whose terms are supplied by one of its parties is a relation with a direction. The derivation requires no dualism and posits no opposed substances. It requires only that the relation be generative, which is what a relational theory of a living world must in any case maintain.

Relational ontology does not deliver harmony. It delivers the necessity of power, and with it the necessity of governing it.

Harmony, in the relational theory as Qin develops it, lies at the foundation. In the theory developed here it lies at most at the end, as an accomplishment, and the greater part of the work consists in specifying the conditions under which the accomplishment is possible. The disagreement is not about whether relations come first, on which the two theories agree, and it is not a disagreement about the value of harmony, which both hold in view. It concerns what a relational ontology is entitled to conclude once relations are granted priority, and this paper’s answer is that harmony is among the things such an ontology must earn rather than among the things it may assume.

Debts to Others

The materials of the derivation are not new, and it would be idle to pretend otherwise.

That generation produces difference, and that difference is prior to identity rather than derivative from it, belongs to Deleuze. That individuation is a process rather than a state, and that the individual is a phase of a system rather than its atom, belongs to Simondon. That power is exercised rather than possessed, that it is productive as well as repressive, and that it is distributed through a field rather than concentrated at a summit, belongs to Foucault. That a living system maintains itself through structural coupling with a medium it does not command belongs to Maturana and Varela, and to Luhmann after them.

The increment claimed here is not the ontology. It is the joint.

The pieces lie in different hands. Eglash supplies the normative claim, that value ought to return to those who generate it, and has no occasion to ask why asymmetry should arise at all, since in the domains he studies it is simply given. Deleuze and Foucault supply the ontology of difference and emergence and offer no ought, which is no failing in them, since they were not writing normative theory. Qin supplies the relational premise and takes harmony as its basic state. What is attempted here is the joint, and the joint alters what generative justice is. It ceases to be a moral preference among the available preferences, and becomes a response to a structural fact that resists removal.

The difference this makes is worth stating exactly. Were asymmetry contingent, a corruption to be cleaned away, the right response to it would be its elimination, and generative justice would be one proposal among others for what to do in the meantime. If asymmetry is a concomitant of generation, then it can be shaped and is unlikely to be eliminated, and the question of the shape it takes is no longer one question among others. It becomes the question.

Objections

That this is pan-powerism. If a generative relation produces power, it may be said, then the concept has been inflated until it discriminates nothing.

The reply concedes the premise and denies the conclusion. The existence of power is pervasive in generative systems, and at the level of existence the concept does little work. This is conceded without reluctance. The discriminating force lies in the form rather than the existence, and the following section supplies the criterion by which forms are told apart. A theory that says only that power is everywhere has said little. A theory that says power is pervasive and that here is how to distinguish the power that builds from the power that consumes has said what needed saying, and it needed the first clause in order to say the second.

That the derivation is merely verbal. The chain is analytic, it may be said, and analytic chains discover nothing.

The reply concedes the character of the argument and disputes the inference from it. The chain is analytic. Its value lies in removing a primitive rather than in discovering a fact, and the removal of a primitive changes what can be asked. A theory that treats power as given is poorly placed to ask what kind of power it is, having no account of what power is a kind of. A theory that derives power from generation can ask whether a given power serves the generation from which it came, and that question organises what follows.

That joint settlement is the ordinary case. The terms of coordination, it may be said, are commonly negotiated rather than imposed, and treaties, conventions, and standing bodies of law are the evidence.

The reply is to accept the evidence and dispute what it shows. That terms are negotiated establishes that the parties took part in supplying them. It does not establish that they took part symmetrically. A negotiation proceeds within a language, upon a precedent, in a forum, and at a table whose shape was settled before anyone sat down, and each of these was supplied by somebody. The question is not whether the weaker party spoke. It is whether the terms in which it was obliged to speak were of its own making. Joint settlement is therefore not the refutation of the account but an instance of the thing to be explained, and the account predicts what is in fact observed: that negotiated outcomes track, with tolerable fidelity, the distribution of the capacity to furnish the terms in which negotiation is conducted.

That states are not generative processes. The derivation applies, it may be said, to processes of production, learning, and differentiation, and it is an open question whether states are such processes.

The reply is that the question, though empirical, is not a difficult one. States produce technology, knowledge, institutions, culture, and capability; they learn; they differentiate internally and from one another; and the differences so produced are the differences that enter into the coordination which constitutes international politics. What the derivation requires of a state is that it generate, and states generate.

§4 Two Kinds of Asymmetry

If asymmetry resists removal, then the question that matters is less how much of it there is than what kind it is, and a theory with one category of asymmetry is poorly placed to ask. Realism has one category. Asymmetry appears within it as a differential of capability, and a differential of capability is read as a threat, since the theory offers little else for it to be read as. This section supplies the second category, and with it a criterion for telling the two apart.

The Distinction

An asymmetry is generative when the parties differ, when the difference is complementary, and when the relation raises the capacity of each party to generate.

An asymmetry is extractive when the parties differ and when one of them employs the difference to lower the other’s capacity to generate for itself.

The two are difficult to distinguish at the level of the scalar. Both are asymmetries. Both may present identical ratios of capability, identical distributions of resources, identical flows of goods across the relation. Little that can be read off the magnitudes tells them apart, and a theory that reads magnitudes alone will therefore misclassify systematically, treating the first as a danger to be corrected and correcting away the condition of its own prosperity.

The Criterion

An asymmetry is generative or extractive according to the net change, across a closed loop of interaction, in each party’s capacity to generate independently.

The weight of the criterion rests on the final word, and the word must be given its force.

The signature of extraction is not that the weaker party produces less. It is that the weaker party produces more, and produces under conditions that the stronger party supplies. Output rises; independence falls. The relation is experienced as beneficial, defended as beneficial, and shows as beneficial on a ledger that counts what was produced rather than what could be produced were the relation to end. Dependency, in its developed forms, is manufactured through provision rather than through deprivation, which is why it is so seldom recognised by those it is practised upon and so readily defended by those who practise it.

The canonical contrast is that between the transfer of a production line and the transfer of a design capability. The first raises output and lowers independence, since what has been installed can be operated but is less easily modified, extended, or replaced, and its operation continues so long as the supplying party continues to supply. The second raises independence, since what has been transmitted is the capacity to make the thing rather than the thing itself, and the capacity, once transmitted, is not readily recoverable by the party that gave it. The two transfers may be of identical monetary value and may look alike in any account that money can keep.

The Criterion Is Observable

The criterion is behavioural and does not require the tracing of value flows, which is in any case impracticable.

One observes the degree of dependence: what proportion of the recipient’s productive activity requires inputs, licences, standards, or authorisations that the counterparty controls. One observes survival under severance: what the recipient can continue to do when supply is withdrawn, and how quickly its capacity degrades. One observes the rate of subsequent independent innovation: whether the recipient, having received, goes on to make things on its own account.

And one observes a further test, which is the discriminating one.

A loop is generative if it produces, in the recipient, capacities the donor did not design.

The test is sharp because generosity does not satisfy it. A donor may give a great deal, may give it freely, and may give it with every intention of benefit, and the recipient may nonetheless be able to do only what the donor equipped it to do. Where the recipient begins to produce what the donor did not anticipate and could not have specified, something has passed between them that was not in the transfer, and it is that passage which the criterion is meant to catch.

Three Kinds of Loop

The criterion yields three cases rather than two, and the third is absent from the existing normative literature.

The extractive loop. At least one party’s independent generative capacity falls across the circuit. This is the case that generative justice was designed to identify, and it identifies it correctly.

The sustaining loop. Neither party’s independent generative capacity changes across the circuit. Little is taken and little is built. The relation persists, the accounts balance, and no one is wronged.

The generative loop. Both parties’ independent generative capacity rises across the circuit. The relation produces a surplus that neither party held before and that neither holds against the other.

The middle case does not arise in the domains from which generative justice was built, and the reason is instructive. The braiding shop and the cooperative are generative by construction: they exist in order that something be made, and a relation among makers that produced nothing would have dissolved before it could be studied. Where the generative character of the relation may be taken for granted, a criterion that detects extraction is sufficient, since extraction is then the only way the relation can fail.

International relations may not be taken for granted in this way. A relation between states may persist for a century while producing little, sustained by inertia, by the cost of exit, or by the interest of both parties in the appearance of a relation. Such a relation is formally clean and substantively inert. Value has not been extracted from those who generated it, and the loop has closed; but little was generated, and the closure of the loop is the closure of an account in which few entries were made. The absence of theft does not establish the condition of a relation, and a framework carried onto terrain where inert relations are common requires a third category, which on its home terrain it had no occasion to need.

An Incompatibility

There follows a conclusion about realism which is derivable and which is therefore stated without polemic.

Realism’s criterion of success, that the system does not collapse and that capabilities are distributed symmetrically, stands in tension with the signature of a generative loop, which is pointwise asymmetry sustained over time.

A system that fully achieves the realist objective is unlikely, by the criterion advanced here, to be generative.

No claim is made that realists desire to suppress generation, and none is needed. The point is that the success condition of the one theory approaches the failure condition of the other, and that the tension follows from the definitions rather than from anyone’s motives.

The instance is one that realism claims with confidence. Two blocs held for four decades in a balance of nearly perfect symmetry, a system that did not collapse, a distribution of capabilities that each side laboured continuously to keep level: on the realist criterion, a success, and it has been so recorded. On the criterion advanced here it is a sustaining loop at best, and over much of its extent an extractive one, in which each party expended its generative capacity upon the single domain in which the other had locked it, and the capacity of each to generate much else declined for as long as the lock held. That the system did not collapse is true, and it is among the least interesting things that can be said about it.

Limitations

The criterion has three weaknesses and they are named here rather than buried.

Measurement. Independent generative capacity is observable in principle and difficult in practice. The proxies proposed above are proxies, and each can be gamed by a party that knows it is being measured by them. No fully satisfactory operationalisation is offered here.

The boundary of the loop. The criterion is defined across a closed circuit, and the determination of where a circuit closes is itself a judgment. In many cases it is a contested one, and the party that gets to draw the boundary of the loop is in a position not unrelated to the position that gets to set the rate. This is not resolved here.

Aggregation. The gravest of the three weaknesses is aggregation. A state’s generative capacity is an aggregate over an internal structure of great heterogeneity, and it may rise in the aggregate while falling for a region, a sector, or a class. A transfer that raises a state’s capacity to generate may do so by concentrating the capacity in a metropolitan centre that is itself extractive with respect to its own periphery.

The criterion, applied at the level of the state, may therefore certify as generative a loop that conceals extraction one level down. A framework that stopped here would risk complicity at the level it declined to examine. What is required is a multi-level treatment, and this paper does not provide one.

§5 The Heterogeneity Problem

The argument now arrives at the operation with which it is principally concerned. The preceding sections have derived asymmetry from generation, shown that it resists removal that leaves generation standing, and supplied a criterion by which the asymmetry that builds may be told from the asymmetry that consumes. The criterion asks whether, across a closed circuit of the relation, each party’s capacity to generate independently has risen or fallen. It presupposes that the circuit can be closed, and that the closing can be inspected. What follows is a difficulty with that presupposition, and it is a difficulty that arises on the terrain of international politics and not on the terrain from which the criterion was drawn.

The Cargo of International Loops

The relations that carry value between states carry it across registers. A state supplies technology and receives recognition. A state supplies capital and receives legitimacy. A state supplies a security guarantee and receives basing rights and strategic alignment. A state supplies an apology and receives, or fails to receive, the closure of a historical grievance. A state supplies market access and receives votes in a multilateral body. Each of these is an ordinary transaction of international life, and in each of them the thing that goes out and the thing that comes back belong to different kinds.

They are, in the relevant sense, incommensurable. There is no evident substance of which technology and recognition are both quantities, and no measure that stands behind a security guarantee and a basing right and expresses each as so much of a common stuff. The claim is not about the present state of measurement, as though a sufficiently refined instrument would eventually supply the missing scale. It is a claim about the values themselves. They are not produced from a common source, they do not enter the world as differing quantities of one thing, and aggregation is unlikely to recover a unity that was not there to begin with.

The cases from which the theory of generative justice was developed are not ordinarily of this structure. In the workshop, in the cooperative, in the community of makers, what the labourer creates returns to the labourer, and it returns as income, as skill, as ownership, as time. The return may be adequate or inadequate, timely or delayed, but it is a return of the same kind as the contribution, and its adequacy is therefore a matter that can be inspected. One counts what went out and one counts what came back, and the accounts are in the same currency because they are accounts of the same thing. This is not a defect in the theory. It is a property of the domain in which the theory was worked out, and it is a property that the international domain does not share.

Closure as Judgment

The consequence is that a question which is empirical in the homogeneous case becomes something else in the heterogeneous one.

Where return is homogeneous, the question whether a loop has closed is a question of fact. Did the value generated by the labour come back to the labourer? One inspects, and one finds that it did or that it did not. The finding may be difficult, the records may be poor, the causal chains may be long, but the difficulty is of a kind that better inspection could in principle overcome. The question has an answer, and the answer does not turn on who is asking.

Where return is heterogeneous, this is no longer so. For the loop to close, something must be made to count as repayment for something of a different kind. Recognition must count as repayment for technology. A basing right must count as repayment for a guarantee. The word that carries the weight is count, and what it names is an operation, not a discovery. Little in technology and little in recognition determines how much of the second discharges an obligation incurred by the first. The two stand in no ratio, since a ratio requires a shared dimension along which it is taken.

In the heterogeneous case, the statement that a loop has closed is not a report of a fact. It is a judgment, and a judgment requires terms on which it is made. Those terms are not supplied by the values judged.

This is the operation that will be called conversion throughout what follows, and the terms on which it is performed will be called the rate of conversion. The vocabulary is deliberately plain. What is at issue is the everyday operation by which an account containing entries of different kinds is declared square, and the question is where the terms of that declaration come from.

Three Evasions

The argument requires that conversion be difficult to avoid. Were there a way to answer the question of closure in the heterogeneous case without positing a rate, the rate would be an artefact of one method among others and would carry little weight of its own. Three such ways present themselves, and each must be closed. They are set out in the order of their initial plausibility.

The First Evasion, Monetisation

The first proposal is that heterogeneity is only apparent, because everything can be priced. Technology has a value in money, recognition has a value in money, and once each has been assigned its price the account can be inspected in the ordinary way. The heterogeneity that seemed to obstruct the audit turns out to be a heterogeneity of appearance only, dissolved by the universal solvent that modern economies have in any case already applied to nearly everything else.

There are three replies, and they are cumulative.

The first is that the goods with which international politics is principally concerned resist pricing, and that their resistance is not incidental to what they are. Recognition, dignity, sovereignty, and historical justice are political goods, and they are political in the measure that they resist pricing. A party that accepts a price for its recognition has, in accepting, conceded much of what recognition was to secure, since what recognition secures is standing as a party rather than as a quantity, and a party that may be bought out at a price is a quantity. The refusal of a price is better understood as constitutive of the good in question than as obstinacy in the face of a valuation that could in principle be made. Espeland and Stevens, whose account of these matters is developed at length below, give the general form of this observation: there are values whose incommensurability is not a residue of imperfect measurement but a condition of the identity of those who hold them.

The second reply is more damaging, because it shows the evasion to be an instance of the thing evaded. To price recognition is to establish a rate at which recognition settles against money. The pricing is less an alternative to conversion than an instance of it. It is a conversion, performed at a rate, and it differs from other conversions chiefly in that the register into which everything is translated is a monetary one, and is so familiar that the translation passes unnoticed. The proposal therefore does not answer the question of where the rate comes from. It answers it in one particular way, by fixing the rate in money, and it then declines to defend the fixing on the ground that it is natural. This is the founding move of a great many settlements, and the ease with which it is made is precisely why it must be refused.

The third reply is empirical, and it is worth stating because the first two are conceptual and the reader may suspect that conceptual scruple is being purchased at the cost of the facts. If monetisation closed the loop, then monetised settlements of political debts would settle them. They routinely do not. Reparations are offered and refused. Compensation is accepted and the grievance persists, and persists in a form that the acceptance of the compensation has not touched and sometimes has aggravated. The offer of money for a wrong is regularly experienced as a second wrong, and the experience is better read as perception than as confusion: the rate at which money settles against the wrong was set by the party that did the wrong. A theory unable to accommodate this would miss a great deal of what it is meant to explain.

The Second Evasion, Categorial Separation

The second proposal is that conversion may be refused. Let each register be audited on its own terms. Let technology settle against technology and recognition against recognition, and let no account be permitted to cross the line between them. If the registers never meet, no rate is ever required.

The reply is in three parts, and the first is simply that the world is not so arranged. The loops of international politics are cross-register by constitution rather than by corruption. Development assistance is exchanged for votes in multilateral bodies. Security guarantees are exchanged for bases and for alignment. Market access is exchanged for political accommodation. These are less degenerate cases in which a purer practice has been contaminated than the practice itself. A theory of international value circulation that begins by excluding them has not idealised its object so much as abandoned it.

The second part of the reply is a reductio. Suppose the separation were enforced, and each register audited alone. Then in nearly every bilateral relation nearly every register would be found unbalanced, because the balancing item lies in another register. Technology flows one way and is not repaid in technology. Recognition flows the other and is not repaid in recognition. On the proposed accounting, most relations in international politics would be extractive in most registers, and the conclusion is absurd. The absurdity indicts the accounting rather than the parties, and it indicts it at the point where the accounting refused to convert.

The third part is the shortest. The separation does not answer the question of settlement. It forbids the question. To say that registers may not settle against one another is less to have found a way of settling them than to have declared that settlement does not occur, in a world where it occurs continually and constitutes the substance of the relations under study. An evasion that succeeds by denying its explanandum is better described as a refusal.

The Third Evasion, Subjective Satisfaction

The third proposal is the most modest of the three and, for that reason, the most tempting. Let the parties decide. The loop has closed when both parties agree that it has closed. There is no need for an external rate, because the rate is whatever the parties in the event accept, and a settlement that both parties accept is by that fact a settlement.

The reply is that this criterion consults, as its evidence, the very thing it is supposed to evaluate.

Consider what determines whether a party accepts a settlement. It accepts when what it has received registers, within its own scheme of valuation, as adequate to what it gave. But its scheme of valuation is not independent of the relation. It has been formed within the relation, over time, under conditions in which the terms of adequacy were being continually rehearsed by the stronger party and continually inhabited by the weaker one. The satisfaction of the weaker party is therefore an effect of the rate, and to use it as the test of the rate is to validate the rate by its own consequence. The circularity is not a subtle one, and refinement of the wording is unlikely to avoid it, since it lies in the structure of the situation rather than in the phrasing of the test.

The colonial case is the paradigm, and it is why appeal to sincerity is unlikely to salvage the criterion. The colonised subject who is satisfied with the settlement is not thereby shown to have received a just settlement. It is among the accomplishments of a long relation that satisfaction can be produced, and a theory of justice whose highest court is satisfaction has, at the decisive moment, little to say. What it has is a report on the success of the relation in producing the disposition that vindicates it.

It should be added that this evasion, if accepted, would make the mechanism examined in a later section invisible by construction. The question of how unfavourable rates come to be accepted is the question this paper regards as central. A criterion that treats acceptance as the terminus of inquiry has determined in advance that the question will go unasked.

The Propositions

The three evasions are those that present themselves, and each has failed. Monetisation performs the conversion it claims to avoid. Separation forbids the question it claims to answer. Satisfaction consults the effect of the rate as the test of the rate. What remains may be stated directly.

Proposition 5.1 (Necessity of Conversion). Where values of heterogeneous registers circulate in a relational system, the judgment that a loop has closed presupposes a rate of conversion among those registers.

Proposition 5.2 (Positedness of the Rate). The rate of conversion is not derivable from the values it converts, since values of heterogeneous registers stand in no ratio to one another. The rate is therefore posited rather than discovered.

From these there follows a corollary that will govern the remainder of the paper, and that costs it a great deal.

Corollary 5.3 (No Correct Rate). There is no correct rate of conversion. A critique that presupposes one has imported a common measure unavailable for the values at issue.

The expense should be stated at once, and it will be stated again when it falls due. Corollary 5.3 removes the classical vocabulary of exploitation from the hands of this paper. That vocabulary works by comparison to a true rate. The labour theory of value supplies such a rate, and it is what permits the theory of unequal exchange to say that an exchange conducted at a certain ratio of prices is unequal, for inequality is a relation to a measure and requires that a measure exist. A critique that denies the measure is in no position to say that the rate is wrong. It has surrendered the sense in which a rate could be wrong.

This paper accepts that surrender deliberately, on the ground that the alternative is worse. A theory of international value that retains a common measure retains it at the cost of the values that constitute international politics, which are not produced by labour and are not quantities of any substance. A theory equipped to say that coffee was underpriced and unable to say anything at all about whether a guarantee was repaid, whether an apology was owed, or whether recognition was purchased too cheaply has bought its rigour by narrowing its object until the object no longer contains what it was supposed to explain. The surrender of the measure is the price of speaking about the things that international politics is about, and the question of what remains of critique once the measure is gone is the question to which this paper’s normative section is devoted.

The Limits of the Argument

The limits of the argument should be stated plainly, since the section carries the weight of the paper and an overstatement here would collapse it.

It has not been shown that conversion is illegitimate. Conversions occur, they are difficult to avoid, and little said here counts against them. It has not been shown that a particular rate is wrong, and by Corollary 5.3 no such thing could be shown, since there is no rate against which a rate might be found wanting.

What has been shown is narrower and sufficient. In a settlement across registers a rate is present; it is posited; and it is not derivable from the things it settles. From which it follows that in such a settlement someone has done the positing. The question of who, and of what it is to occupy the position from which positing is done, is the subject of the sections that follow.

§6 Settlement Where Refusal Is Foreclosed

The preceding section established that every settlement across registers presupposes a posited rate, and closed with the observation that someone must have done the positing. It is tempting to proceed directly to the naming of a power, and the temptation should be resisted, since the power has been named. This section therefore begins by stating what it inherits, and it inherits a great deal.

The Inheritance

The operation this paper has called conversion is the operation that sociology calls commensuration, and it has been studied under that name for close to three decades. Espeland and Stevens define commensuration as the expression or measurement of characteristics normally represented by different units according to a common metric, and observe that it transforms qualities into quantities and difference into magnitude. They devote a section of their account explicitly to the question of how commensuration refracts power relations, and they answer that it is a mode of power, that its power is both strategic and constitutive, that it reconstructs relations of authority and creates new political entities and establishes new interpretive frameworks, and that it is emphatically not the neutral technical process its advocates claim. They observe that its operation becomes invisible through routinisation, so that the labour it requires and the assumptions surrounding it are forgotten, and that commensuration constitutes what it purports to measure. They read Marx as holding that labour is the great commensurator, and that value commensurates by abstracting away the distinctiveness of the particular forms of work. And they observe, in a passage that bears directly on the present subject, that politics understood as the art of compromise is itself a broad instance of commensuration, so that interest-group bargaining presupposes the very operation whose terms it is unable to examine.

This paper does not re-derive the foregoing. It stands on it. The claim that the setting of terms of settlement is a mode of power, that this power is constitutive rather than merely strategic, and that its operation is concealed by its routineness, belongs to the sociology of commensuration and is credited to it without reservation.

What this paper adds lies elsewhere, and it can be located precisely by attending to the direction in which the sociological literature travels.

The Boundary and the Interior

The cases from which the sociology of commensuration is built are cases of encroachment. A metric advances into a sphere that had been held apart from it, and the political question that arises is whether the advance shall be permitted. The economist attempts to synthesise a price for an activity that had not been priced. Advocates of comparable worth attempt to bring domestic labour within the metric of wages. Legal scholars attempt to establish, or to resist, a market in children. In each case a mode of valuation is pressing upon a region where another mode has prevailed, and the drama is a drama of the border.

Resistance, in such cases, tends to take a standard form. It is the claim of incommensurability. Following Raz, Espeland and Stevens name a strong version of this claim the constitutive incommensurable: a category of value whose incommensurability is bound up with the identity of those who hold it, so that to permit the trade-off is to become someone else. The Yavapai, whose ancestral land was threatened by a dam, held that land as a constitutive incommensurable, and the models by which bureaucrats evaluated the dam required that the land be made commensurate with the other components of the decision. The requirement was not merely inconvenient. It was a contradiction of the values in question and of the identity that those values defined. The parents who decline to think of their children as substitutable, the homemakers who refuse to have their lives priced against occupational attainment, and the movements that decline to bring their commitments to the bargaining table are all making the same move, at different scales and with different degrees of success. They are saying: this is not the kind of thing that may be traded against that.

Espeland and Stevens even name the diplomatic instance, and they read it in exactly this way. The territorial commitments of the Israeli and the Palestinian peoples, they observe, have confounded countless efforts to commensurate territorial interests at diplomatic bargaining tables. The case appears in their account as a case of refusal, and the refusal is what makes it interesting to them.

The sociology of commensuration is a theory of the boundary. Its subject is the advance of a metric and the refusal that resists it, and the political possibility on which its account of resistance rests is the possibility of saying: not this, not here, not on those terms.

International politics, in the greater part of its ordinary operation, takes place in the interior. The boundary was crossed long ago, and it is not readily recrossed. The parties are already being settled against, continually, and the question before them is not whether to permit the metric but on whose terms the metric will run.

The Foreclosure of Refusal

The claim of the section may now be stated. The refusal that the sociological literature documents requires a position from which it can be sustained, and a state that lacks an exit from the relation is poorly placed to occupy it.

Consider what refusal accomplishes in the sociological cases. The Yavapai, in declaring their land incommensurable, do not thereby prevent the dam. They may lose. What the declaration secures, whether or not it prevails, is that the loss is registered as a loss rather than as a settlement. They have not been paid for the land, since they have refused the terms on which payment would have been payment. The wrong remains a wrong, and remains available to be named as one, since the accounting that would have closed it was declined.

A state is rarely placed to do this. It may declare its sovereignty incommensurable, and it will be settled against notwithstanding. It may hold that no quantity of investment discharges the obligation incurred by a colonisation, and the ledger will nevertheless be drawn up, the investment entered on one side and the extraction on the other, and the account declared square by parties who did not require its assent. The declaration is heard. The account is closed anyway.

The reason is not that states are weaker than communities, which they are not. It is that the refusal available to the Yavapai is a refusal to participate, and participation is what a state is least able to decline. To refuse settlement is to withdraw from the field within which settlements are made, and to withdraw from that field is to become, not a state that has declined to be settled against, but a state that does not register as a party. The mechanism is taken up in the following section. Here it is enough to record the structural fact.

Proposition 6.1 (Foreclosure of Refusal). Where a party cannot withdraw from a relation without ceasing to count as a party, the claim of incommensurability is unavailable to it as a form of resistance. Commensuration then ceases to be a contested boundary and becomes a constitutive condition of the relation.

The international case is then not one further instance of a phenomenon that sociology has described, but its limit, at which the phenomenon changes character. Where refusal is available, commensuration is a border under dispute, and the political question concerns the location of the border. Where refusal is foreclosed, the border is not in dispute, and the political question concerns the terms on which the interior is governed.

The foreclosure of refusal has a consequence for the normative argument developed later, and it is worth marking here, since it converts a proposal into a derivation. Where the strong form of resistance is unavailable, what remains is contestation of the rate rather than refusal of it. The normative condition advanced below is less chosen from among alternatives for its attractiveness than left standing when the others fall.

Against the Faces of Power

The commensuration literature has not placed its object against the standard taxonomy of power. The placement is worth performing, both to locate the operation and to forestall the objection that it is a known one under a new description. Each of the received forms of power is examined in turn, and the operation of settlement is found to lie beside each of them rather than within any.

Dahl’s coercive power gets one party to do what it would not otherwise do. The setting of a rate compels no action. It operates upon the ledger rather than upon the will. A party may act freely throughout, choose each of its own moves, and find at the end that its account has been settled at a rate it did not set.

The second face of power controls the agenda and excludes items from it. Here nothing is excluded. A grievance may be raised, tabled, and entered into the account. What the rate determines is less what may be said than what a thing that has been said is worth against another thing. A grievance may be fully voiced, fully admitted, and fully entered, and settled at a rate the aggrieved did not set.

Lukes’s third face shapes wants, so that grievance does not arise. The operation examined here works upon equivalences rather than upon desires, and the two come apart in the ordinary case. A party may want what it would have wanted under conditions of undistorted formation, may articulate the want with clarity, may press it with vigour, and may find that it has been discharged by an offer of a different kind, since the terms on which that kind settles against this kind were not its to establish. The third face prevents the complaint. The rate prices it.

Keohane and Nye come close here, and the point deserves care. Their insight is that the asymmetry of a dependence is itself a resource of power, and that this power is detachable from military capability. But the power they describe is a scalar: a matter of who can better tolerate the severance of the relation, measured along the single dimension of that tolerance. Bargaining leverage presupposes that the matters in dispute have been rendered comparable enough to be bargained over, and the setting of the rate is what renders them so. The rate is therefore prior to bargaining rather than a species of it, and the framework that reduces the question to leverage has performed a conversion of its own without remarking upon it.

Gramscian hegemony explains how an order comes to be consented to. The operation in question here is the one by which a good of one kind comes to count as settlement for a good of another. Legitimation is a consequence of that operation, and the theory of hegemony describes the consequence while leaving the operation unsupplied.

Bourdieu lies nearer than the rest, and requires careful handling. His symbolic power naturalises an arbitrary order and secures its acceptance through misrecognition. But misrecognition is a relation to an arrangement already in place, and naturalisation presupposes something to naturalise. Bourdieu explains with great power why the rate goes unquestioned. He leaves unexplained how there came to be a rate, and the two questions are distinct.

Nye’s soft power is a stock of attractiveness, a resource held by a party and expended in the ordinary way, and it belongs therefore to the scalar tradition this paper has been at pains to leave behind. What is at issue here is less a resource than a position: the position from which equivalences are declared.

The Colony with a Positive Ledger

The framework shows its use where the rival frameworks are silent, and there is a case in which they are.

The central operation of colonial rule was not plunder. Plunder is visible, calculable, and protestable, and it was frequently contested and occasionally curtailed. The central operation was the unilateral positing of the rate, and it is preserved in the sentence that colonial administration composed for itself: we gave civilisation, railways, law, and medicine; you returned resources, labour, and land; the account is square.

The word that carries the argument is square. It presupposes a rate at which railways settle against sovereignty, and medicine against land. The rate was set by the metropole, and the language in which its fairness might have been disputed was set by the metropole also, so that the colonised party wishing to object was obliged to conduct its objection in terms of what a railway is worth, which is the question that had already been answered.

Now consider the case in which the material ledger runs positive. There are colonial relations in which net capital flowed inward, in which railways were built and schools founded and mortality reduced, and in which an honest accounting of material flows shows the colony to have received more than it sent.

The rival frameworks are poorly placed to find the injustice.

Generative justice does not find it, since value flowed toward the generators rather than away, and the loop, materially, closed. The theory of unequal exchange does not find it, since it detects exploitation through a negative balance and requires the labour theory of value to establish that the balance is negative; where the balance is positive the theory has little to grip. And the sociology of commensuration sees the power at work and declines, by its own methodological commitment, to call it unjust: it observes, with studied even-handedness, that commensuration may enlarge decision-making or legitimate a preordained decision, that it may be manipulated by elites or may constrain them, and it stops there.

This framework finds the injustice and locates it. It lies not in the flow, which ran the other way, nor in the unfairness of the rate, since by Corollary 5.3 there is no fair rate against which this one might be found wanting.

It is in the foreclosure of any position from which the rate could be reopened.

Realism as a Regime of Settlement

One further application is required before the argument proceeds, and it concerns the discipline within which the argument is offered.

Realism converts recognition, justice, identity, culture, and legitimacy into a single scalar, which it calls capability, and it holds that these things matter to the conduct of states in the measure, and only in the measure, that they can be so converted. This is a commensuration, and it is performed at a rate.

The claim of realism to be a science rests upon its not acknowledging as much. It presents the conversion as a description of what moves states: in the end, it says, only power counts. But that everything else can be settled against power is less an observation about the world than a stipulation about the terms of the account, and it is the founding stipulation of the field. The conversion is prior to the findings, and the findings do not test it, since they are recorded in the ledger the conversion established.

There is a consequence, and it explains something long noticed and never accounted for. When realism, in its prescriptive mode, informs a state that its claims to recognition count for little unless backed by capability, it is less reporting a regularity than enforcing a rate. The tendency of realist prescription to bring about the conditions it describes has been remarked upon for as long as the doctrine has existed, and has generally been attributed to the psychology of statesmen or to the dynamics of the security dilemma. The mechanism proposed here is simpler. A theory that supplies the terms on which accounts are settled, and is believed by those who settle them, need persuade no one of anything in order to make the world it describes. It need only be used.

§7 The Ground of the Rate

This section is offered at a lower level of confidence than those preceding it. It advances a mechanism, and it advances it because there is an explanandum which the available mechanisms do not explain, and not because it has been demonstrated. The status of the argument is stated at the end of the section and should be borne in mind throughout it.

The Question Left Open

If the rate is posited rather than discovered, and if it therefore has no authority in the nature of the things it settles, why is it not simply contested into symmetry? Two parties confront one another across an account whose terms are arbitrary. Each has an interest in terms favourable to itself. Why do the terms of one of them prevail?

The available answers are insufficient, and each falls short for a reason that is instructive.

Coercion is insufficient because rates persist where coercion has lapsed. The formal sovereignty of the postcolonial state is complete, and its accounts continue to be settled on terms established elsewhere. Were the rate held in place by force, its persistence after the withdrawal of the force would require explanation, and little is forthcoming.

Bargaining power is insufficient because it presupposes what is to be explained. To bargain over a settlement is to bargain within a set of equivalences already established, and the question is where the equivalences came from. An account that grounds the rate in bargaining power has placed the effect before the cause.

Information is insufficient because rates persist under conditions of full information. Nobody is deceived. The weaker party is quite capable of stating that the terms are arbitrary, that they were set by the stronger, and that its acceptance of them serves the stronger’s interest. It states this, often with great precision, and accepts the terms.

The Locus of Judgment

The proposal begins from a feature of the operation left so far unexamined. An equivalence is less a property of the two things it relates than a judgment, and a judgment is made from somewhere.

To say that a basing right counts as repayment for a security guarantee is less to report a fact about basing rights and guarantees than to say that, within some scheme of accounting, the entry on the one side discharges the entry on the other. The scheme is somewhere. It is occupied, and the party that occupies it is the party for whom things count.

Proposition 7.1. The power to posit the rate of conversion is structurally linked to occupying the position relative to which things count. Whoever occupies that position sets the rate, not by compelling assent to it, but by being the one for whom the account is an account.

The structure has been worked out with care in the psychoanalytic vocabulary, in which the position for which a subject’s productions count, and from which they receive their value as productions, is the position of the Other. Little in the argument requires acceptance of the clinical apparatus this vocabulary carries with it, and a reader who declines the apparatus may read what follows as a claim about the pragmatics of recognition and will lose little.

The Acceptance of Unfavourable Rates

The mechanism earns its place by explaining a phenomenon that realism handles poorly and that has been evident for as long as anyone has looked.

Realism predicts that a state possessing an exit option and confronting an unfavourable settlement will exit. States routinely do not. They remain in relations whose terms they can articulate as unjust, whose arbitrariness they can demonstrate, and from which the material cost of withdrawal would be bearable. The behaviour is neither marginal nor confined to the weak. It is the ordinary conduct of the international system, and a theory that predicts otherwise owes an account of why the prediction so regularly fails.

The proposal is this.

Exit from the relation is exit from the field within which one counts.

To withdraw is to become, not a state that has declined an unfavourable settlement, but a state that does not register as a party. The relation is less the site at which the account is kept than the condition of being the kind of thing that has an account.

The weaker party therefore accepts an unfavourable rate in order to remain within the accounting. It purchases its legibility with the surplus.

This closes a gap left open by the preceding section. The foreclosure of refusal was there recorded as a structural fact about the international system, and a structural fact recorded without a mechanism is an observation awaiting explanation. The mechanism is now supplied. Refusal is foreclosed because it costs the refuser its standing as an actor, and standing as an actor is less one good among the goods at stake in the settlement than the condition of having anything at stake.

It follows that the weaker party’s acceptance is better read as a purchase than as weakness of will, miscalculation, or false consciousness. What it purchases is the one thing that holding out would not obtain.

A Testable Consequence

The proposal has an implication that discriminates it from the alternatives and that could come out against it.

If acceptance of an unfavourable rate is bought in exchange for standing, then acceptance should track the party’s dependence upon the relation for recognition as an actor, and should not track, or should track only weakly, its material dependence and the availability of material alternatives. Two states with equal material dependence and equal alternatives should diverge in their willingness to accept unfavourable settlements according to how much of their standing as parties is constituted by the relation in question.

If, on the contrary, acceptance is fully accounted for by material dependence, the mechanism proposed here is superfluous and should be discarded. Little else in the paper depends upon it.

The Status of the Proposal

This section proposes a mechanism. It does not demonstrate one.

The claim that the position from which equivalences are declared is structurally the position for which things count is advanced as a hypothesis with explanatory reach. Its principal support is that it accounts for a phenomenon the alternatives do not: the acceptance of unfavourable rates where exit is materially available and where nobody has been deceived. That is a reason to entertain it. It is not a proof, and is not offered as one.

§8 Contestability

The Surrendered Ground

Corollary 5.3 denied that there is a correct rate of conversion. The denial was required, since the values that circulate in international relations stand in no ratio, and measurement is unlikely to recover a ratio that was not there. But the denial removes the ground on which critique has classically stood.

The classical form of critique against an exchange is that the exchange was conducted at the wrong rate. This is what the theory of unequal exchange says, and it can say it because it possesses a measure: labour time supplies the true value against which the price is found wanting, and the finding is a finding, not a preference. Take the measure away and the sentence loses its meaning. One is not placed to say that a rate is unjust where there is no rate against which it might be compared, and this paper has said that there is none.

A critical theory that has abandoned the standard by which it criticises is in serious difficulty, and the reader will have noticed. The difficulty is the hardest moment of the argument, and it is not smoothed over.

The remainder of the section attempts a recovery. It should be judged as an attempt.

The Residue of Resistance

Begin from what the sociology of commensuration has already established about resistance.

Where a metric advances upon a sphere that has been held apart from it, resistance tends to take a characteristic form. It is the claim of incommensurability: the assertion that this is not the kind of thing that may be traded against that. The Yavapai make it of their land. Homemakers make it of their lives against the metric of occupational attainment. Movements make it of their commitments when they decline to bring them to the bargaining table. It is the strong form of refusal, and where it can be sustained it is the apt one, since it does not contest the terms of an accounting but declines to have the accounting.

The strong form requires a position from which it can be held. The party must be able to walk away and remain what it was.

A party that cannot withdraw without ceasing to count is poorly placed to make that claim. It will be settled against whether it consents or not, and its declaration that the account may not be drawn is heard, and the account is drawn.

Contestability is what remains of resistance when refusal has been foreclosed.

It is not a weaker substitute selected for convenience, nor a proceduralist retreat from a substantive standard that might have been maintained. It is what is left standing, and its necessity is derived rather than preferred. A party that cannot say this is not that kind of thing may still say not at that rate, and not by your reckoning alone.

The Condition

Proposition 8.1 (Contestability). Where no true rate of conversion is available, the normative condition of a settlement is not the justice of the rate but its contestability: whether the party against whom the rate is applied possesses standing, language, and forum in which to dispute it.

The three components are each required, and the second is the demanding one.

Standing is the recognised entitlement to dispute the rate, and it is to be distinguished from the entitlement to petition against its application. A party that may plead for lenience in a particular case, but may not question the terms on which cases of that kind are decided, has been given the appearance of the thing and not the thing.

Language is the availability of terms in which the dispute may be conducted that are not wholly supplied by the party that set the rate. This is the condition that the preceding sections have shown to be usually violated, and violated in the manner that makes the violation hardest to see. The colonised party wishing to dispute the rate at which railways settle against sovereignty must conduct the dispute in terms of what a railway is worth, and the question of what a railway is worth was answered when the rate was set. It is less that the objection is suppressed than that the objection, in order to be made, must be made in the coin that is under dispute.

Forum is the existence of a site at which the dispute is heard and at which it can have effect. A dispute that can be raised nowhere has been permitted in the manner in which the impossible is permitted.

Against the Charge of Proceduralism

The objection will be made that this reduces justice to procedure, and that a settlement may be perfectly contestable and perfectly iniquitous.

The reply is not that procedure suffices for justice. It is that where no substantive standard is available, the property of being reopenable is the one property a rate can have that is not itself a posit. The features of a rate’s content were put there by somebody. Its openness to being reopened was not.

A rate that cannot be reopened is not thereby shown to be wrong, since there is little that would show it wrong. It is thereby placed beyond the reach of the parties it binds.

That, rather than its content, is what constitutes the injustice.

The proposal is a substantial revision of the normative tradition and should be stated as one. It removes symmetry from the conditions of justice. A settlement may be radically asymmetrical, transferring far more one way than the other by any accounting one cares to construct, and may nonetheless satisfy the condition proposed here, provided the terms on which it was reached remain open to those it binds. And a settlement may be scrupulously balanced and fail the condition, if the balance was struck on terms that one party alone established and the other has neither standing, language, nor forum in which to question.

Consequences

Three follow directly.

The first is that a formally closed loop with an incontestable rate is extractive, whatever the material ledger shows. This is the operational form of the diagnosis offered for the colony with a positive balance, and it is what permits that diagnosis to be made at all.

The second is that symmetry has been removed from the conditions of justice, as stated above.

The third concerns institutional design, and is developed further below. The question to be put to an institution is not whether it distributes fairly. It is whether the rates of conversion embedded within it can be reopened by the parties they bind.

Limitations

Three weaknesses, and the third is grave.

The threshold. How much contestability suffices? The proposal does not say. Standing, language, and forum admit of degrees, and no account is given of how much of each is enough. The gap is real, and it is not filled here.

Capture. A forum may exist in which disputes are heard and reliably lost, and in which the existence of the forum serves to legitimate the loss. The condition, as stated, is satisfied by such a forum, and it should not be. What is required is a further specification distinguishing a forum that hears from one that absorbs, and it is not supplied.

Whether it is strong enough to condemn. Suppose that a colonial rate had been contestable; suppose that the colonised had possessed standing, language, and forum; suppose the rate had been contested and upheld. Would the settlement then have been just?

The answer this paper gives is that it would have been legitimate in a sense in which it was not, and that the distinction is real rather than verbal. The colonised party that has been heard and has lost stands in a different relation to its loss than the party that was not permitted to speak. But the answer is uncomfortable, and the discomfort should be admitted rather than argued away. A reader who finds that the framework, at this point, purchases its consistency too dearly is not making a mistake. He has identified the price.

§9 Toward a Formal Representation

This section is a research direction rather than a result. The substantive claims of the paper are stated and defended in natural language, and none depends upon what follows. It is included because the structure of the problem is unusually well matched to an existing mathematical apparatus, and it would be strange to notice this and say nothing.

The Motivation

The operations that have organised the argument share a character. Values that stand in no ratio must nonetheless circulate in a single loop. Comparison must be made without a global standard, since no such standard is available for the goods in question. And the comparison that matters is less that of one party with another than that of a party with itself, before and after a circuit of the relation.

These are the operations for which the mathematics of connections was developed. A connection is the object permitting comparison across the points of a space that lacks a global frame. It says how a quantity defined at one point is to be carried to another, and does so without requiring that quantities at different points belong to a common set. The problem this paper has posed is close to the same problem in another language, and the coincidence is worth recording.

The Sketch

Let the base space be the graph of relations among actors.

Let the fibre above each actor be that actor’s space of generative capacities. The fibres are not compared across actors and are not required to be commensurable with one another. This is less a limitation of the construction than the point of it. The framework neither requires nor permits us to say what a technology is worth against a recognition.

Let a connection be given by the transport of capacity along an edge: an interaction, occurring between two parties, alters what each is able to generate.

Then the holonomy of a closed circuit is the net alteration in a party’s capacity after the capacity has been carried around the loop and returned to the party from which it started. The comparison is local. It compares a party with itself in its own fibre, and requires no cross-party metric.

The sign of the holonomy is the criterion of the fourth section. Positive: each party can now generate independently what it could not before. Zero: the sustaining loop. Negative: extraction.

The attraction of the construction is precise, and it is worth stating what it buys. The incommensurability that defeats the theory of unequal exchange, and that obliged this paper to surrender the classical vocabulary of exploitation, is here structurally accommodated rather than assumed away. A common measure of value is not needed in order to ask, of each party, in its own fibre, whether it can now generate more than it could.

The Difficulties

Four, and they are not small.

The ordering of generative capacity. The criterion of the fourth section requires more and less. Where the space of generative capacities is not ordered, the sign of the holonomy is undefined and the construction says little. That generative capacity admits of an ordering is not obvious, and it is not established here. Capacity may be only partially ordered, in which case some pairs of loops would be incomparable and the criterion would apply to fewer cases than has been supposed.

The specification of the connection. To say that an interaction transports capacity is at present a redescription rather than a specification. Until the transport rule is given, the formalism has little content, and a formalism without content is a picture.

The place of conversion. The fourth difficulty carries the widest consequences. If the settlement of heterogeneous values is the central object of the paper, then the formalism must find a place for it. The natural suggestion is that the rate of conversion is the connection: it is the rule by which what is given in one fibre is registered in another, and it is precisely the sort of thing a connection is, being neither a property of the source nor a property of the destination but a rule for carrying between them.

Were that identification made rigorous, this section would become the spine of the paper rather than its appendix, and the identification of the posited rate with a chosen connection would supply a formal sense in which a rate is arbitrary: a bundle carries no canonical connection, and the choice of one is a choice.

It has not been made rigorous here, and the reader should not be left with a contrary impression.

The risk of ornament. A formalism that licenses no inference otherwise unavailable is decoration, and decoration in a work of political theory is worse than nothing, since it purchases an appearance of rigour at the cost of the thing. The construction earns its place if it permits a conclusion the argument in words does not.

Status

This paper does not demonstrate that the formalism earns its place. It states the correspondence and the difficulties, and leaves the matter open. It is the author’s principal object of further work.

§10 Implications for Governance

The normative condition advanced above has consequences for institutional design, and the consequences are not the ones the reform literature ordinarily pursues. This section states them, and then states the problem that stands in their way and that this paper does not solve.

The Design Question

If a settlement is just in the measure that its rate remains open to those it binds, then the question to be put to an institution is not whether it distributes fairly.

Can the rates of conversion embedded in this institution be reopened by the parties they bind?

Three Observations

The first is that international institutions characteristically fix their rates of conversion in their constitutive documents, and characteristically make those provisions among the least revisable parts of themselves. Voting weights, veto structures, the classes of claim that are admissible and those that are not, the terms on which a contribution in one register is credited against an obligation in another: these are less rules of procedure than rates, and they are entrenched, and entrenched more deeply than most of what the institution contains.

The second is that the standard reform debate concerns redistribution within a fixed rate. More seats, more votes, more voice, a larger share of an allocation whose terms of allocation are not in question. The prior question is who may reopen the rate, and it is seldom asked, in part because the rate does not present itself as a rate. It presents itself as the constitution of the body, which is to say, as the thing within which questions are asked rather than as a thing that may be asked about.

The third is a tentative design principle, offered as such.

An institution should place its rates of conversion at a lower level of entrenchment than its distributive rules, and not, as is universal, at a higher one.

The ground is that distributive error is visible and correctable, whereas rate error is neither. A party that has received too small a share can say so, and in terms the institution supplies. A party settled against at a rate it did not set is poorly placed to say so in those terms, since the terms are what it wishes to dispute. The provision hardest to see is therefore the provision that ought to be easiest to change, and the institutions now in existence have the matter the other way about.

The Problem of Exit

The theory of generative justice relies, at the decisive point, upon the agency of those who generate. They may leave. They may form a cooperative, keep what they make, and reorganise the loop on terms of their own. Much of the value of the theory rests upon this possibility, and where the possibility obtains the theory is powerful.

In the international system, exit is foreclosed, and the argument of the seventh section proposes that it is foreclosed in a deeper sense than the material one. A state that withdraws becomes, not a state that has declined an unfavourable settlement, but a state that does not register as a party, and it loses thereby less the benefit of the relation than the standing in which benefits are had.

Then who reorganises the loop?

The answer is unlikely to be the parties themselves, which is the answer that generative justice gives. This paper has no other to offer.

The question of who reorganises the loop is the principal open problem of the argument, and it is not a small one. It bears upon whether the theory of generative justice survives transposition to the international level, or whether it remains a theory of communities in which the door is open and becomes, when carried to a system in which the door is shut, a description of a hope rather than a programme.

The Problem of Levels

A second problem, recorded above and restated here as a problem of governance.

A loop may be generative at the level of the state and extractive one level below. A transfer that raises a state’s capacity to generate may raise it by concentrating that capacity in a metropolitan centre which stands to its own periphery as the donor stands to it. An institution that certified the loop as generative would then have certified an extraction it had not examined.

The framework must eventually be a multi-level one. It is not one here, and until it is, it should be applied with the knowledge that it may prove complicit at the level it declines to examine.

§11 Cases as Tests

The cases that follow are tests rather than illustrations. Each is chosen because it discriminates between this framework and a named rival, and each is stated so that it might come out against the framework. Where a case would fail to discriminate, it is omitted, however vivid.

The Production Line and the Design Capability

This tests the criterion of the fourth section.

Two transfers of technology of equal monetary value are made to comparable recipients. In the first the recipient acquires the capacity to operate a process; in the second it acquires the capacity to design one. The prediction is that the two will diverge in the rate of subsequent independent innovation and in survival under severance of supply, and will do so even where the material ledgers are identical and the money transferred the same.

What would falsify it. Should recipients of process transfers show the same rate of subsequent unanticipated innovation as recipients of design transfers, the distinction between output and independent capacity is measuring little, and the criterion should be abandoned.

The Colony with a Positive Ledger

This tests the sixth section against three rivals at once, and it is the paper’s principal case.

The case is chosen precisely because the material accounting runs in the colony’s favour: net capital inward, infrastructure built, schools founded, mortality reduced. Generative justice should find no injustice here, since value flowed toward the generators. Unequal exchange should find none, since it detects exploitation through a negative balance. The sociology of commensuration finds power and, by its own commitment, declines to call it unjust.

The framework advanced here finds the injustice and locates it in the foreclosure of any position from which the rate might have been reopened.

What would falsify it. Should it be shown that the injustice in such cases is fully accounted for by a negative flow once the accounting has been done properly, the rate is doing little work in the explanation and the framework is redundant. Much of the paper’s claim to increment rests upon this case, and the case is offered in that knowledge.

Assistance in Exchange for Votes

This tests the fifth section, and it tests two of the three evasions in a single case.

Development assistance is exchanged for support in multilateral bodies. The loop resists separation into registers, since its two sides are constitutively of different kinds, and the second evasion therefore fails within it. And the loop is routinely monetised in the literature, the value of the assistance being priced and the value of the vote estimated by the price the assistance commanded, so that the first evasion is performed openly and may be watched failing: the price of the vote is derived from the assistance, the adequacy of the assistance is then assessed against the price of the vote, and the circle is closed without remark.

The Gift Refused and the Gift Accepted

This tests the sixth and seventh sections.

Two diplomatic gifts, structurally identical in every material respect: the same value, the same occasion, the same relation between the parties. The first is refused as insulting. The second is accepted.

On a material accounting the two cases are alike, and no ledger distinguishes them. The difference lies in whether the conversion was accepted, which is to say, in whether the recipient consented to the terms on which a thing of one kind was to count as settlement for a thing of another.

What would falsify it. Should the acceptance and the refusal leave no persisting consequences, so that the relation proceeds alike whichever occurred, the settlement was epiphenomenal and the sixth section is idle.

Unfavourable Rates Under Available Exit

This tests the mechanism of the seventh section, and it is the case on which that section stakes itself.

The prediction is that a state’s acceptance of an unfavourable settlement will track its dependence upon the relation for recognition as an actor, and will track only weakly its material dependence and the availability of material alternatives. Two states materially alike, with alternatives alike, should differ in what they will accept according to how much of their standing as parties is constituted by the relation in question.

What would falsify it. Should acceptance prove fully explained by material dependence, the mechanism is superfluous and should be discarded. Little else in the paper depends upon it.

The Cold War as a Sustaining Loop

This tests the incompatibility asserted in the fourth section, and it is included because it is the case that realism claims most confidently as its own.

By the realist criterion the case is a success: a system that did not collapse, a distribution of capabilities that each party laboured continuously to keep level, a balance sustained across four decades. The framework advanced here predicts that over the same period the independent generative capacity of each bloc, in every domain other than the one in which the two were locked, declined or stagnated relative to a counterfactual in which the lock was absent.

The test is hard and the counterfactual difficult, and this is not concealed. It is included because a framework with little to say about the case realism most confidently claims would have declined the argument at the point where it is joined.

§12 Conclusion

The Argument

Power is not a primitive. It follows from generative differentiation, and deriving it rather than receiving it changes what may be asked about it. A theory that takes power as given is largely confined to asking how much of it there is. A theory that derives power from generation can ask whether a given power serves the generation from which it came.

Asymmetry is not injustice, and it resists removal that leaves generation standing. This is Eglash’s insight, and what has been added here is the ontological ground it lacked. To which is joined a consequence that the relational theories of world politics have not drawn from the premise they share with this paper: relational ontology does not deliver harmony. It delivers the necessity of power, and with it the necessity of governing it.

In a loop that carries heterogeneous values, closure is a judgment rather than a fact. Something must be made to count as repayment for something of another kind, and the terms on which it is so made are not supplied by the values themselves. The rate is posited, and there is no correct one.

And where a party cannot withdraw from a relation without ceasing to count as a party, the strong form of resistance is unavailable to it. It is not placed to decline to be settled against. Commensuration then ceases to be a contested boundary and becomes a constitutive condition of the relation.

The international case is less one further instance of the phenomenon that the sociology of commensuration has described than its limit, at which the phenomenon changes character.

Realism, finally, is a regime of settlement. It converts recognition, justice, identity, and culture into a single scalar, and presents the conversion as a description of what moves states, where it is better understood as a stipulation about the terms on which their accounts are kept.

Proposals Not Established

That the position from which equivalences are declared is structurally the position for which things count. The mechanism explains what the alternatives do not, and that is a reason to entertain it rather than a demonstration.

That the normative condition, in the absence of a true rate, is the contestability of the rate. The proposal has a threshold problem and a capture problem, and it is uncomfortable in a case that ought to be easy.

That the mathematics of connections is more than a suggestive analogy for the structure of the problem. It has not been shown that the formalism licenses any inference the argument in words does not.

Questions Left Open

Who reorganises the loop when exit is foreclosed. The theory of generative justice answers that the generators do it themselves; in the international system they are seldom placed to, and this paper has no other answer. It is the central unsolved problem, and it bears upon whether generative justice survives its transposition.

Whether a loop generative at one level conceals extraction at the level below. The framework must eventually be multi-level. It is not one here.

How much contestability suffices. No threshold is given.

The Argument Compressed

Generative justice requires that value return to those who generate it. Where value is heterogeneous, return must be settled across registers, and settlement requires a rate that cannot be derived from what it settles.

The sociology of commensuration has shown that the setting of such a rate is a mode of power, and that the ordinary resistance to it is to refuse the metric outright.

This paper concerns the case in which that refusal is unavailable: in which a party is not placed to decline to be settled against, since to decline is to cease to count.

There, the condition of justice is not that the rate be fair, for there is no fair rate, but that it remain open to being reopened.

Position Within the Programme

Two remarks, offered without inflation.

Within the wider programme of which this paper forms part, the argument is the political instance of a structure worked out elsewhere. The settlement of heterogeneous values across a relation from which neither party can withdraw is formally close to the settlement of heterogeneous gifts within an intimacy, and the correspondence is a theorem rather than a metaphor. It is not developed here.

Within the programme’s governance branch, the paper supplies a design criterion, which is the contestability of rates, and states the obstacle that must be overcome if the criterion is to be more than a wish. The obstacle is the foreclosure of exit, and it has not been overcome.

The paper closes, as it must, on the thing it does not do. It has argued that a party which cannot leave a relation is poorly placed to refuse the terms on which it is settled against, and that its remaining recourse is to contest them. It has not shown how such a party acquires the standing, the language, and the forum in which contestation would be possible. That is the question, and it is left open because it is open.

References

Amin, Samir. Unequal Development: An Essay on the Social Formations of Peripheral Capitalism. New York: Monthly Review Press, 1976.

Anderson, Elizabeth. Value in Ethics and Economics. Cambridge, MA: Harvard University Press, 1993.

Bachrach, Peter, and Morton S. Baratz. “Two Faces of Power.” American Political Science Review 56, no. 4 (1962): 947–952.

Bourdieu, Pierre. Language and Symbolic Power. Cambridge, MA: Harvard University Press, 1991.

Broome, André, and Joel Quirk. “Governing the World at a Distance: The Practice of Global Benchmarking.” Review of International Studies 41, no. 5 (2015): 819–841.

Chakrabarty, Dipesh. Provincializing Europe: Postcolonial Thought and Historical Difference. Princeton: Princeton University Press, 2000.

Cox, Robert W. “Social Forces, States and World Orders: Beyond International Relations Theory.” Millennium 10, no. 2 (1981): 126–155.

Dahl, Robert A. “The Concept of Power.” Behavioral Science 2, no. 3 (1957): 201–215.

Deleuze, Gilles. Difference and Repetition. Translated by Paul Patton. New York: Columbia University Press, 1994 [1968].

Eglash, Ron. “An Introduction to Generative Justice.” Teknokultura 13, no. 2 (2016): 369–404.

Emmanuel, Arghiri. Unequal Exchange: A Study of the Imperialism of Trade. New York: Monthly Review Press, 1972.

Espeland, Wendy Nelson. The Struggle for Water: Politics, Rationality, and Identity in the American Southwest. Chicago: University of Chicago Press, 1998.

Espeland, Wendy Nelson, and Mitchell L. Stevens. “Commensuration as a Social Process.” Annual Review of Sociology 24 (1998): 313–343.

Fanon, Frantz. The Wretched of the Earth. Translated by Constance Farrington. New York: Grove Press, 1963 [1961].

Foucault, Michel. Discipline and Punish: The Birth of the Prison. Translated by Alan Sheridan. New York: Pantheon, 1977.

Gilpin, Robert. War and Change in World Politics. Cambridge: Cambridge University Press, 1981.

Hornborg, Alf. “Unequal Exchange Is Not Primarily about Monetary Value.” Nature Communications 16 (2025).

Jackson, Patrick Thaddeus, and Daniel H. Nexon. “Relations Before States: Substance, Process and the Study of World Politics.” European Journal of International Relations 5, no. 3 (1999): 291–332.

Keohane, Robert O., and Joseph S. Nye. Power and Interdependence: World Politics in Transition. Boston: Little, Brown, 1977.

Kindleberger, Charles P. The World in Depression, 1929–1939. Berkeley: University of California Press, 1973.

Krasner, Stephen D. “State Power and the Structure of International Trade.” World Politics 28, no. 3 (1976): 317–347.

Lukes, Steven. Power: A Radical View. London: Macmillan, 1974.

Maturana, Humberto R., and Francisco J. Varela. Autopoiesis and Cognition: The Realization of the Living. Dordrecht: Reidel, 1980.

Mauss, Marcel. The Gift: The Form and Reason for Exchange in Archaic Societies. Translated by W. D. Halls. New York: Norton, 1990 [1925].

Mearsheimer, John J. The Tragedy of Great Power Politics. New York: Norton, 2001.

Mennicken, Andrea, and Wendy Nelson Espeland. “What’s New with Numbers? Sociological Approaches to the Study of Quantification.” Annual Review of Sociology 45 (2019): 223–245.

Morgenthau, Hans J. Politics among Nations: The Struggle for Power and Peace. New York: Knopf, 1948.

Nye, Joseph S. Soft Power: The Means to Success in World Politics. New York: Public Affairs, 2004.

Organski, A. F. K. World Politics. New York: Knopf, 1958.

Ostrom, Elinor. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press, 1990.

Polanyi, Karl. The Great Transformation: The Political and Economic Origins of Our Time. Boston: Beacon Press, 1944.

Porter, Theodore M. Trust in Numbers: The Pursuit of Objectivity in Science and Public Life. Princeton: Princeton University Press, 1995.

Qin, Yaqing. A Relational Theory of World Politics. Cambridge: Cambridge University Press, 2018.

Qin, Yaqing. “The Zhongyong Dialectic: A Bridge into the Relational World.” The Chinese Journal of International Politics 17, no. 2 (2024): 206–221.

Radin, Margaret Jane. Contested Commodities. Cambridge, MA: Harvard University Press, 1996.

Raz, Joseph. The Morality of Freedom. Oxford: Clarendon Press, 1986.

Sen, Amartya. Development as Freedom. Oxford: Oxford University Press, 1999.

Simondon, Gilbert. Individuation in Light of Notions of Form and Information. Translated by Taylor Adkins. Minneapolis: University of Minnesota Press, 2020 [1958].

Spivak, Gayatri Chakravorty. “Can the Subaltern Speak?” In Marxism and the Interpretation of Culture, edited by Cary Nelson and Lawrence Grossberg, 271–313. Urbana: University of Illinois Press, 1988.

Sunstein, Cass R. “Incommensurability and Valuation in Law.” Michigan Law Review 92, no. 4 (1994): 779–861.

Walt, Stephen M. The Origins of Alliances. Ithaca: Cornell University Press, 1987.

Waltz, Kenneth N. Theory of International Politics. Reading, MA: Addison-Wesley, 1979.

Wendt, Alexander. “Anarchy Is What States Make of It: The Social Construction of Power Politics.” International Organization 46, no. 2 (1992): 391–425.

Zelizer, Viviana A. Pricing the Priceless Child: The Changing Social Value of Children. New York: Basic Books, 1985.

---

中文

权力与价值在国际关系中的初步讨论

一种生成性关系的视角

黄万宏（huangwanhong@serendip.ngo）

摘要

国际政治的诸理论测量权力、汇总权力，并追问其分布会带来什么后果。它们很少追问权力何以存在。权力是可以被推导的。一个进行生成的过程必定要制造差异，因为一种令世界与其原初状态无从区分的生成实则一无所生；它所差异化出来的东西，包含了供给协调所据以进行之条款的能力；而一段其条款由其中一方供给的关系，是一段带有方向的关系。因此，不对称是”有所生成”的关系所处的条件，而非对这种关系的败坏，并且它抵抗那种在移除它之后仍让生成得以存续的清除。于是正义的问题不可能是如何消除不对称，转而成为：不对称对它所身处的那段关系做了什么。

在国际政治中回答这个问题，会撞上一个现有框架无法应对的障碍。一个国家供给技术而接受承认，供给保障而接受基地权。在这样的诸善之间并无比率，而为了让账目结清，某物必须被迫算作对另一种物的偿付。这项操作需要一个比率，而这个比率并非由它所结算的那些价值给出。它是被设定的，且不存在一个正确的比率。三条回避之路被封闭。货币化施行了它宣称要避免的那项转换；范畴分离禁止了它宣称要回答的那个问题；主观满意则以比率的效果作为比率的检验。

在不同类之物之间设定一种共同尺度乃是权力的一种样式，这一点属于测度化的社会学，并且在此是被继承而非被主张的。但那套文献理论化的是边界：一种尺度向一个被隔离于其外的领域挺进，而抵抗采取彻底拒绝该尺度的形式。拒绝预设了一个人可以从中退出并保持其原样的位置。一个若不退出便不再算作一方的国家，没有这样的位置，并且它会不顾其反对而被结算。在拒绝被封闭之处，测度化不再是一条有争议的边界，而成为关系的一个构成性条件，因而国际的情形并非该社会学现象的又一个实例，而是它的极限，在此极限上该现象改变了其性质。

由此有两个推论。物质账簿呈正数的殖民地，一个生成性正义、不平等交换与测度化社会学同样对之沉默的情形，得到了诊断：不义既不在于流向，也不在于比率的不公平，而在于任何可以从中重开该比率之位置的被封闭。而一项结算的规范性条件，被定位于其比率的可争议性，这既非一种程序主义的退却，也是拒绝被封闭之后残留下来的抵抗。该论证放弃了剥削的古典词汇，那套词汇需要一种国际政治之诸善所不容许的共同尺度。本文以它并不回答的那个问题作结，即：谁来重组一个无一方可以从中退出的回路。

关键词： 权力；不对称；测度化；生成性正义；均势。

AI 使用声明： 本文的观念、主张与理论框架源自作者。AI 工具被用作起草、经验与文献来源的定位与核实以及修订过程中的辅助。每一个观念的出处均已由作者核实，作者对内容承担全部责任。

§1 引言

均势理论把权力当作一种存量。它测量权力，将其汇总到一个体系的各单位之上，并追问其分布方式会带来什么后果。它追问一种分布是否稳定，一种集中是否危险，一种转移是否趋向战争。它很少追问何以有任何权力可供分布。

这一遗漏的后果超出它发生之处。一种将权力当作原初项来接受的理论，很难说清权力是何种事物，因为它对”权力是何物之一种”没有任何交代，而一种对种类保持沉默的理论，被局限于数量的记法之内。它能说这里多、那里少，多者具威胁而少者易受害，且谨慎的政策倾向于二者的均等化。它难以言说的，是一次给定的权力施用究竟建造了某物还是消耗了某物，因为这是一个关于一段关系之性质的问题，而该理论手上只有那些量值。

权力是可以被推导的。一个进行生成的过程必定要制造差异，因为一种令世界与其原初状态无从区分的生成实则一无所生。它所差异化出来的东西，包含了供给协调所据以进行之条款的能力，因为条款的供给本身就是一项生产性行为，与被生产出来的一切其他事物一样落于同一差异化之下。而一段其条款由其中一方供给的关系，是一段带有方向的关系。权力就是那个方向的名字。

不对称在这一说法上表现为”有所生成”的关系所处的条件，而非对关系的败坏，并且它抵抗那种在移除它之后仍让生成得以存续的清除。由此立即有两个推论。现实主义长久坚持的悲剧，既不需要对支配的胃口，也不需要无政府状态：它会在任何生产出任何东西的体系中出现。而那个提议废除不对称的纲领是自我挫败的，其方式是结构性的而非实践性的，因为在它移除差异的程度上，它会移除生成所在于其中的那种差异化。

于是正义的问题不可能是如何消除不对称，转而成为：不对称对它所身处的那段关系做了什么。有些不对称提升双方各自独立生成的能力；另一些则提升较弱一方的产出，同时降低其独立生产的能力，而这正是依赖通常被制造出来的方式，通过供给而非通过剥夺。这一区分对于一种读取量值的理论是不可见的，而正是这一区分，本文其余的一切都系于其上。

因此，追问价值是否回到那些生成它的人那里，成为该问的正确问题，而在国际政治中它立即撞上一个障碍。一个国家供给技术而接受承认。它供给资本而接受正当性。它供给安全保障而接受基地权。它供给一份道歉而接受，或未能接受，一桩宿怨的了结。出去的东西与回来的东西属于不同的记法，而在这些记法之间没有共同的实质，也没有比率。为使这样一个回路结清，某物必须被迫算作对另一种物的偿付。这项操作需要条款，而这些条款并非由价值本身供给，因为不同记法的价值彼此之间不处于任何量值的关系中。因此条款是被设定的，而一旦看清它们是被设定的，就必须追问是谁设定了它们。

这一障碍是普遍的，且它逐一击败了现有的框架。现实主义不应对它，因为它在困难得以出现之前，就已经把全部困难转换成了单一的标量。不平等交换理论应对它，并以一种共同尺度即劳动时间来回答它，而这种尺度对于保障、承认和道歉是不可得的。而生成性正义理论，它已确立价值理应回到那些生成它的人那里，却是在回报同质的领域中建立的：一门手艺所生产的价值回到工匠那里，作为收入、技艺、所有权或时间，而人们可以查验它是否回来了。在回来之物与给出之物同类的场合，账目以一种货币记录，无须任何转换比率。一旦被带到诸记法互异的地带，”价值必须回来”这一要求就成了一项在有人说出何者算数之前无法核查的要求。

设定不同类之物据以彼此结算之条款者是谁，这是一个古老的问题，而其中许多阵地已被占据。埃斯佩兰与史蒂文斯在 1998 年确立，在不同类之物之间设定一种共同尺度乃是权力的一种样式，该权力是构成性的而非仅仅是策略性的，它通过例行化而变得不可见，且政治性讨价还价预设了它。后殖民批评确立，回到被殖民者那里的东西是以殖民者的语言回来的。建构主义确立，意义先于利益，且制度是沉积下来的实践。世界政治的诸关系理论与本文的论证共享一个前提，即关系先于实体。而埃格拉什确立，价值理应回到那些生成它的人那里，这正是本文将其带到无法核查之地带的那个主张。所欠是巨大的，通篇均予承认，而随后展开的论证，正是这些债务偿清之后所剩下的东西。

测度化社会学是一种关于边界的理论。它的诸案例是侵入的案例，其中一种尺度向一个被隔离于其外的领域挺进，而政治的戏剧就是这一挺进是否将被允许的戏剧。抵抗，在那套文献中，往往采取一种特有的形式，即不可通约性的主张：断言这不是可以拿来与那个交换的那类事物。亚瓦帕伊人对他们的祖传土地作出此断言。持家者对他们的生活作出此断言。运动在拒绝走上谈判桌时对他们的承诺作出此断言。这是拒绝的强形式，而在它能够被维持之处它就是恰当的那一形式，因为它并不争执一项计账的条款，而是拒绝根本进行这项计账。但强形式需要一个可以据以坚持它的位置。该方必须能够走开并保持其原样。

一个国家很少处于能够如此的位置。它可以宣布其主权不可通约，而它会不顾此而被结算。它可以主张任何数额的投资都不能了结一次殖民所招致的义务，而账簿仍会被拉出来，投资记于一侧，攫取记于另一侧，且由无须其同意的诸方宣布账目结清。宣布被听见了。账目无论如何被结清了。原因不是国家比共同体更弱，它们并不更弱。原因是拒绝结算就是从结算得以做成的场域中退出，而从那个场域中退出，就是成为，不是一个已拒绝被结算的国家，而是一个不登记为一方的国家。

在拒绝被封闭之处，测度化不再是一条有争议的边界，而成为关系的一个构成性条件。于是国际的情形并非社会学所描述过的那个现象的又一实例，而是它的极限，在此极限上该现象改变了其性质。

该论证付出了某种代价，与其在事情进行过程中被发现，不如在开端处就予以声明。如果结算的比率是被设定而非被发现的，那么就不存在一个正确的比率，而一种因某项结算是在错误比率上达成而谴责它的批判，已经擅自借用了它并不拥有的尺度。这移除了剥削的古典词汇，那套词汇通过与一个真实价值相比较而运作，并要求一个真实价值是可得的。不平等交换理论在劳动时间中拥有这样一种尺度。对于构成国际政治的那些善，没有可比的尺度可得，它们并非由劳动生产，也不是任何实质的量，因此必须另寻他据。

所提议的据是可争议性。在真实比率不可得之处，一个比率所能拥有的、其本身并非一项设定的唯一属性，就是它对被它所约束者重新开启的开放性。一个比率之内容的诸特征是由某人放进去的。而它可以被质问这一点却不是。而既然抵抗的强形式，即彻底拒绝该尺度，对一个不能退出的一方是被封闭的，那么对条款的争议就不是一种为了方便而选择的较弱替代品。它是所剩下的东西。

一个推论关乎该论证被提出于其中的这门学科。现实主义把承认、正义、身份、文化与正当性转换成单一的标量，并主张这些东西之关涉于国家的行为，其程度、且大体上仅其程度，在于它们能被如此转换。这本身就是一次跨记法的结算，在某个比率上施行，而现实主义之为国际政治之科学的主张，正建立在其不承认这一点之上。它把这一转换呈现为对何者驱动国家的描述，而它更宜被理解为一项关于账目条款的规定。一切别的东西都能被结算于能力之上，与其说是一项关于世界的观察，不如说是该领域的奠基性设定。

本文按如下方式展开。文献被加以归类，不是按学派，而是按每一传统对不对称所做之事，而该归类定位出一处空白，同时承认填满相邻空白的那些东西。随后从生成推导出权力，并将此推导转而针对现实主义、针对平等主义，以及针对那些与本文论证共享前提、却从中得出”关系根本上是和谐的”这一结论的诸关系理论。接着提供一个用以区分”建造的不对称”与”消耗的不对称”的准则。确立国际价值的异质性，并随之确立一个被设定之比率的不可避免性，其方式是封闭那三条本可用来回避它的回避之路。梳理拒绝被封闭的诸后果，并将其应用于物质账簿呈正数的殖民地，一个对手框架对之沉默的情形。以较低的置信度提议一个机制，用以说明弱者为何接受他们完全能够看出乃是任意的比率。以一个规范性条件替代那个被放弃的条件。勾勒一种形式化表征并陈述其困难。本文以它并不回答的那个问题作结，即：谁来重组一个无一方可以从中退出的回路。

§2 关于不对称的文献

本节并不综述学派。它按各理论对不对称所做之事来归类，而该归类本身就是一个论证。三个问题组织了它。

问一。发生学问题。 不对称从何而来？

问二。规范性问题。 不对称本身是否构成不义？

问三。结算问题。 当所交换的价值属于不同记法时，账目如何结清？

本节的论点是，问三主要在一处得到过回答，即测度化社会学，且在那里所回答的是诸方可以拒绝该尺度的情形。拒绝不可得、因而结算是强制性的那种情形，则鲜受关注。那种情形正是国际政治的寻常情形。

作为给定的不对称

现实主义家族把不对称当作分析的起点，而非当作一件有待解释之事。古典现实主义把它奠基于对支配的胃口；结构现实主义奠基于能力的分布；进攻性现实主义奠基于被不确定意图所加重的无政府状态；威胁均衡奠基于被感知的威胁而非原始能力；权力转移理论奠基于差异化的增长率。霸权稳定论，在它们中间独树一帜，允许能力的集中是有益的。

问一处的困难扎得很深，并非侧重点的问题。华尔兹的能力分布是重新描述了待释项而非解释了它。权力转移理论对差异化增长的诉诸是重申了待释项：不对称之所以出现，是因为事物以不同的速率增长，而以不同速率增长正是不对称之所在。摩根索对人性的诉诸是该传统对发生学的一次持续尝试，且它是一个人类学假说而非结构性假说，这使它很难解释权力在一个其成员是国家而非个人的体系中的涌现，除非借助一个鲜有辩护的类比。

此外，结构现实主义内部还有一个本文认真对待、而该传统尚未解决的困难。华尔兹宣布结构由诸单位之间的位置关系所构成，随后却把结构三要素中的第三个计算为单位属性之和。结构被主张为关系性的，却被计算为实体主义的。本文认真对待那个关系性主张，并将其引向异于华尔兹所得的诸后果。

霸权稳定论值得单独处理，因为它临近本文的论证，且是友非敌。它承认该家族其余成员所不承认者，即能力的一次集中可以强化它所被集中于其中的那个体系，条件是被集中的一方供给公共品而非攫取租金。但它没有提供任何可据以将这两种情形分辨开来的准则。它在事后、并借由结果来区分它们，而结果是一份记录而非一条准则。下一节所提出的准则，正是作为霸权稳定论一直缺失的那件东西献给它的。

作为可治理的不对称

自由主义家族把不对称当作一种有待中和的危险。制度降低交易成本并供给信息；机制稳定预期；民主和平把危险定位于政体类型而非差别本身。

基欧汉与奈必须被认真对待，因为《权力与相互依赖》临近本文的关切。他们的洞见是，相互依赖关系中的不对称是该关系内部的一种权力资源，而不仅仅是该关系的一个条件，且这种权力是可与军事能力相分离的。这一洞见是正确的，且它是重要的。

它在问三处的局限却是精确的。不对称的相互依赖被理论化为讨价还价的筹码，而讨价还价的筹码是一个哪一方更能容忍关系断裂的问题。这是一个标量，是沿容忍这一单一维度所作的一次测量，且它预设了争议中的诸事项已被弄得足够可比，以至于可以就其讨价还价。它回答了谁握有筹码的问题。它不提出本文所提出的问题：当一方供给安全保障、另一方供给基地权时，谁欠谁，且以何种尺度衡量？该框架把一段异质的关系转换到单一维度之上，却不评说它已施行了一次转换。

问二处的困难属于整个家族。不对称被假定为，几乎不加论证地，某种有待制度去中和之物。它可能是生成的条件而非对生成的威胁，这一点鲜被考虑，而在制度目标为拉平之处，制度的成就趋向于对那本应被分配之物的压制。

作为结构性支配的不对称

不平等交换

不平等交换理论临近本文的核心问题，而两者之间的关系必须被精确陈述，本文才不至于被斥为一次重新发现。

埃马纽埃尔的机制是工资的差别。在资本流动而劳动不流动、且工资由制度决定因而是外生的情况下，边缘以低于其社会价值的价格出售其产品，核心则以高于的价格出售，而剩余被朝核心的方向转移。阿明扩展了这一说法。埃马纽埃尔所谴责的是一个比率，而他所谴责的比率是以劳动时间标价的。

第一个差异是，埃马纽埃尔拥有一种共同尺度而本文没有。劳动时间是那个不变的实质，价格相对于它被发现有所欠缺，而它正是使那次交换能被称为不平等的东西。没有一种尺度，这个概念便难以形成，因为不平等是相对于一种度量的关系，而在没有度量之处，几乎没有什么东西可供不平等地存在于其中。因此本文所面对的困难比埃马纽埃尔的更难，且这不是偶然的：在技术与承认之间，在安全保障与基地权之间，在资本与正当性之间，没有共同的实质，而不仅仅是尚未被发现。

第二个差异随之而来。在真实价值可得之处，人们可以说正确的比率是一回事而实际的比率是另一回事，且这一差额就是剥削。批判瞄准比率对其真值的偏离。本文否认一个真实比率是可得的，因而无法诉诸一个正确的比率。它的规范性支点必须被安置于别处，而这一安置是后面某一节的主题。

共同尺度的放弃是一处脆弱，且它被如此陈述。抛弃该尺度就是抛弃剥削的古典定义。代价是真实的，且它是被有意付出的。

付出它的理由是，劳动价值论不适于国际政治的地带。安全保障、承认、正当性、战略位置与历史性道歉不是由劳动时间生产的，也不是某种实质的量。一种依赖于共同尺度的批判理论，在国际政治最浓厚之处陷入沉默。埃马纽埃尔能告诉我们关于咖啡与纺织品的许多东西。他几乎不能告诉我们马歇尔计划是否得到偿付，什么能偿付一次殖民道歉，或一个国家是否过于昂贵地购买了它的承认。这些不是该领域的边缘案例。它们属于它的核心案例。

应当记录的是，本文的抛弃并不孤单。霍恩博格曾反对生态不平等交换的近期复兴，认为指导埃马纽埃尔与阿明的劳动价值论只能以货币的方式来构想，而从工资的差别推出一次价值的转移，就是把价值实体化为一种在体系的特定点上被创造出来的实质。他追问价值如何能以货币以外的任何度量来表征，并把这一答案的缺席当作该理论的一个困难。本文把这一缺席当作它的资料。霍恩博格在度量的缺乏中发现了对批判的一个问题，本文则在其中发现了一次批判的起点。

两个框架之间还有一处后面会紧要的进一步不对称。埃马纽埃尔需要一个负的物质余额才能侦测出剥削，因而对于资本向内流动的情形几乎无话可说。

葛兰西式政治经济学

考克斯的霸权是强制加上同意，物质化于一个历史集团之中，并通过观念、制度与物质能力而再生产。它接近一种关于凌驾结构的符号权力的理论，且它是一项可观的成就。

差异在于此。考克斯理论化一种特定秩序如何被接受为正当。他不理论化那项操作，凭借该操作一种善成为对另一种善的结算。正当化是那项操作的后果；考克斯极富洞见地描述后果，而把操作留作未供。他能告诉我们战后秩序曾被同意。他较少告诉我们，安全秩序的供给如何成为对货币自主之攫取的一个可接受的回答，也就是说，一种记法中的善如何成为对另一种记法中的善的偿付而被接受。

后殖民理论

后殖民批评已确立本文核心诸节所描述的大部分实质。斯皮瓦克的认识暴力、查克拉巴蒂对欧洲的地方化，以及法农对被殖民主体的说法，会聚于单一的观察：回到被殖民者那里的东西是以殖民者的语言回来的，而回报据以被评估的语言，是作出回报之一方的语言。

这一发现属于后殖民批评。此处所主张者更狭窄，且必须保持狭窄。

这一洞见已被作为文化的与认识论的批判而发展。此处所尝试者是把它重构为一个价值循环体系之内的形式机制，具体而言即一种记法中的善据以了结另一种记法中所招致之债的比率的单方设定。这一重构的目的不是改进该批判，它几乎不需要改进。而是使该批判在一侧可连接于生成性正义理论，在另一侧可连接于制度设计，而文化批判以其自身的语汇是不可如此连接的。这究竟是得是失，留给读者。它在可处理性上是得，在浓厚性上是失，且两者都应说出。

作为被构成的不对称

建构主义确立，意义先于利益，无政府状态是国家造就它的样子，且制度是实践的沉积。英国学派确立，国际社会拥有共同的规范与制度。诸关系理论，在杰克逊与内克松以及秦的著作中，确立关系先于实体。情感转向与实践转向确立，情绪与背景性能力是行为的构成而非其残余。

本文对此几乎无所争执，也不主张其中任何一项。把当前框架定位为”那个解释价值如何涌现的理论”，以对照一种”仅解释价值如何流动”的生成性正义，这一诱惑应被拒绝。那个位置已被占据，且自 1992 年以来一直被占据。

秦的关系理论需要单独处理，因为它把本文纲领所主张者作为其本体论前提而主张，即关系先于实体、行动者是关系中的行动者。它的形而上学内核是 阴 与 阳 的关联关系，它设定其基本状态为和谐，并从该设定推出一切关系根本上是和谐的、且能够被作为和谐来管理。中庸 辩证法供给了认识论图式。

对关系理论的通行反驳，秦已直接应对，追问一种从关系而非从对立的实质出发的理论如何能够辨认主体与客体，以及它随后如何能就二者之间的支配与从属关系给出任何说法。秦的回答是消解这一二分：不可化约之对立的表象是一副二元论透镜的产物，而一种从一开始就不假定任何二元的辩证法，拒绝了困难被提出于其中的那些术语。

本文所给出的回答是相反的那一个，且它是从同一前提之内给出的。它陈述于下一节。

这一类别在问三处的困难是普遍的、结构性的。这些是关于意义之生成的理论而非关于循环的理论。它们很少追问，在一个跨记法的账目必须被结清时谁的意义算数，且它们几乎不就意义的结算功能给出说法，该功能并不从关于意义之生成的说法推出，也不被包含于其中。

测度化社会学

那个已回答问三的传统位于国际关系之外，而它必须被充分给以其应得，因为本文本可为自身主张的大部分东西都属于它。

埃斯佩兰与史蒂文斯把测度化定义为，把通常由不同单位所表征的诸特征依据一种共同度量来表达或测量。它把质转化为量，把差异转化为量值。他们的说法确立了如下诸点，而本文继承而非主张它们。测度化是权力的一种样式，且其权力既是策略性的又是构成性的。它不是其倡导者所主张的那种中立的技术过程，且它重构权威关系、创造新的政治实体、确立新的解释框架。它的运作通过例行化而变得不可见，以至于它所需要的劳作以及围绕它的那些假定被遗忘。它构成它所自称测量之物。马克思在他们的说法中被读作持这样的观点，即劳动是伟大的测度者，价值通过抽掉特定工作的独特性而进行测度，而这正是本文对埃马纽埃尔所给出的读法、因而并非本文的读法。而政治本身，被理解为妥协的艺术，在他们的说法中是测度化的一个宽泛实例，因而利益集团的讨价还价预设了那项它无从审视其条款的操作本身。

在国际关系之内，这一概念是由一扇窄门进入的。布鲁姆与奎克在他们关于全球标杆的说法中运用它，其中测度化作为一条数值转译之链中的一环出现，与简化、外推、实体化和符号判断并列。但那套文献的对象是指标与排名：把规范性价值转译为分数，据此进行远程治理。它的对象不是两方之间一个跨记法账目的结算，而本文所问的问题在那里并不出现。

测度化文献的两处界限界定了本文工作于其中的空间，且它们必须被精确陈述。

第一处是它按设计推却规范性问题，而这一设计在一篇综述文章中是德性，在此处则是局限。测度化，按其说法，可以扩大决策，也可以为一个事先已定的决策提供正当性；它可以被精英操纵，也可以约束他们；它可以武装异见者，也可以生产被规训的主体。这一不偏不倚是严谨而有意的。它意味着该文献没有处于能够说何时一次测度化是攫取性的位置，而本文必须处于该位置。

第二处界限关乎的不是该文献所说的东西，而是它所注视的方向。

测度化社会学据以建立的诸案例是侵入的案例。一种尺度向一个曾被隔离于其外的领域挺进，而政治问题是这一挺进是否将被允许。抵抗于是往往采取一种形式：不可通约性的主张。亚瓦帕伊人对他们的土地说，它不是可以拿来与一座大坝交换的那类事物。埃斯佩兰与史蒂文斯甚至点出那个外交实例，观察到以色列与巴勒斯坦两族人民的领土承诺已在外交谈判桌上挫败了无数次测度领土利益的努力，而该案例在他们的说法中作为一个拒绝的案例出现。

它是一种关于边界的理论。它对抵抗的说法依赖于说出这一话语的可能性：不是这个，不在这里，不按那些条款。

国际政治，在其寻常运作的大部分中，发生于内部。

作为生成之可治理条件的不对称

最后一类在国际关系之内人烟稀少，且大体存在于其外。埃格拉什的生成性正义持不对称不是不义、而价值回路的切断才是不义。奥斯特罗姆供给自治的设计原则而不专题化不对称。森与努斯鲍姆把正义的准则从资源移到能力。莫斯与萨林斯确立礼物的义务性与关系生产性。波兰尼确立并非一切整合都是市场交换、且互惠与再分配运行于不同的逻辑之上。

森的这一步值得承认，因为本文所提出的准则是它的表亲。差异有二。森的能力是个体的一种属性，是一个功能与自由的问题，是一个人能够是什么与能够做什么的问题。此处所运用的生成性能力是关系性的，是一个回路之内某位置的属性而非某节点的属性，且它跨越一个闭合回路来测量而非在一点上测量。问题与其说是一方能够做什么，不如说是：在关系的一次完整回路之后，它是否比先前更能够或更不能够独立生成。

波兰尼是”价值循环于不止一种记法中”这一认识的一位近祖。他没有追问诸记法如何彼此结算，且该问题在他的框架之内并不出现，但正是他使人们能够看见存在着诸记法。

埃格拉什的界限是他的领域的界限，且它不是一处缺陷。他的诸案例是编发店、合作社、造物者的共同体，而在这些之中回报是同质的：由劳动生成的价值回到劳动者那里，作为收入、作为技艺、作为所有权、作为时间。因此他几乎没有机会需要一种转换比率，也几乎没有机会遭遇那个设定一个比率的权力。本文起始于他的领域终止之处。

定位

本文几乎不取代任何东西。它观察到，一个在国际关系中难以回避的问题，在两套并不相遇的文献的交界处被留置未决。

生成性正义理论供给规范，并预设回报是同质的。测度化社会学供给关于异质性的说法，确立结算它乃是权力的一种样式，并按设计推却规范性问题。二者都不触及结算因拒绝已被封闭而成为强制性的那种情形。

而既然没有真实比率可得，本文便无法诉诸一个公平的比率。规范性支点必须被移动。那一困难就是本文。

§3 权力的发生

现实主义理论把权力当作一个原初项。它测量权力、汇总权力、分布权力，并追问其分布的后果，而它很少追问何以有任何权力。这一遗漏不是一处小遗漏。一种对权力从何而来保持沉默的理论，很难说清它是何种事物，而一种对种类保持沉默的理论，大体被局限于追问这里有多少、那里有多少。本节供给现实主义所缺的那个推导，而该推导有些现实主义未必会欢迎、而平等主义会更少欢迎的后果。

推导

从生成到权力的链条分五步。

如果一个过程不生产任何先前不存在的东西，它便不是一个生成的过程。这是词义的问题，除了固定术语之外几乎无关宏旨。

如果所生产之物不引入任何差异，它便未曾被生产。一个什么都不改变的改变不是改变，而一种令体系与其原初状态无从区分的生成实则一无所生。因此生成蕴含差异，是作为它之为生成的条件，而非作为它的一个频繁后果。

一个体系诸部分之间的差异，是允许这些部分彼此处于关系之中的东西。在诸部分无从区分之处，没有什么东西可供一段关系存在于其间。因此关系预设差异，而生成供给差异。

处于关系之中的诸方必须协调，而协调据条款进行。条款并非随关系一同给出。它们必须被供给，而它们的供给是一个进一步的问题，为关系之事实所留置未决。

而在协调之条款由诸方之一方而非共同供给之处，诸方之间的差异获得了一种沿某方向运行的效果。这种带方向的效果就是权力一词所命名者。

条款的来处

第四步与第五步是论证最为暴露之处，而这一暴露应被应对而非略过。可以承认协调需要条款，也承认单方供给的条款产生一种带方向的效果，而其间的那个主张则被否认。为什么条款应由一方供给而非在诸方之间议定？差异本身之中似乎没有任何东西要求如此。两方可以有差异，可以作为平等者议定其协调的条款，而如果那是寻常的情形，那么权力就是讨价还价的一种偶然性而非生成的一个后果，且该推导坍缩为一份对不走运境况的描述。

回应是，条款的供给不是一项站在生成过程之外、有待某种外在于它的安排在诸方之间分配的进一步活动。条款的供给本身就是一项生成性行为。供给一段关系据以进行的条款，就是生产某种先前不存在的东西，而生产它，就是施用那个其不平等分布已被前几步确立的能力本身。

生成生产差异，而它所差异化出来的东西是生成的能力。

协调的条款本身就是某种被生成之物。因此更有能力生成的一方，也更有能力生成条款。

单方设定条件因而不是从外部加于差异之上的一个补充假定。它就是差异，被应用于协调问题本身。

这一点可以作为对反驳者的一个两难而提出。要么协调的条款是被生产的，要么它们是被发现的。如果它们是被发现的，我们必须被告知在何处发现：它们不在诸方之中，因为诸方有差异且各会供给不同的条款；它们也不在关系之中，因为关系正是条款被要求去构成的东西。如果它们是被生产的，那么它们的生产落于已给出的关于生产的说法之下，并受该说法所确立之差异化的支配。因此，那个希望条款在诸方之间作为平等者议定的反驳者，必须主张诸方在生成一切其他东西的能力上不平等，却在生成这一件东西的能力上平等。没有任何理据被给出来支持这一例外，且很难看出可能有什么理据。

由此并不推出更有能力的一方总是供给条款，且主张并非如此。诸能力可以在一个方面不平等而在另一个方面颠倒，而一个在生产上较弱的一方，可以在条款的这一特定生产上较强，那是一个语言、先例、论坛之拥有的问题。所推出者更弱且足够：条款的供给受与一切其他被生成之物相同的差异化支配，而一种对称的供给会要求一种该体系并不以其他方式展现的对称。因此，平等者之间的共同议定不是单方设定条件由以出发的默认状态。它是一项成就，而在它发生之处，是因为某种东西已被做了来生产它。

中项

一个从句承载着这一论证，而缺了它论证便失败。

差异不是权力。

处于对协调之条款的单方控制之下的差异，才是权力。

必须坚持这一从句，因为缺了它推导便滑向”一切差异都是权力”的立场，而一个适用于一切的概念区分不了任何东西。那种滑动是后结构主义传统的一个熟悉的危险，而本文拒绝踏入。两方可以在许多方面有差异并处于一段其条款无一方单独供给的关系之中，而在这样一段关系中差异是丰饶的、且不是权力。把一处差异转换成一种权力者，既不是它的量值也不是它的持久。而是诸方据以协调的条款是由其中一方供给的。

因此差异是权力的潜能而非其现实。这一区分正是允许该理论后来说”有些不对称生成而另一些攫取”的东西。一种把差异等同于支配的理论说不出这一点，因为在这样一种理论上每段关系都会是一次支配，而这个词已不再做工。

读者可能注意到，论证已开始描述本文主要关切的那项操作。不同类之物据以彼此结算的条款，是由一方供给的协调之条款，而此处给出的关于其来处的说法，正是在比率被审视时会再度被需要的那份说法。二者是在不同深度上遇到的同一结构，而后面诸节可被读作对此处已在一般层面上所确立者的展开。

两个命题

命题 3.1（权力的发生）。 权力并不源于资源的占有。它从生成过程之内的关系性差异化中涌现。生成生产差异，而它所差异化出来的东西包含供给协调所据以进行之条款的能力；如此供给的条款给差异以一个方向，而带方向的差异就是权力。因此权力应在一个生成性关系体系中被预期，而它的发生并不取决于任何行动者一方的贪婪、恐惧或意图。

命题 3.2（不对称的不可消除性）。 一个没有差异的体系是一个没有生成的体系。不对称的消除与生成的维持，是不可共同达成的。

三个推论

对现实主义

现实主义长久坚持的悲剧被重新定位，且在被重新定位时被加深。它既不源于人对支配的胃口，如古典现实主义者所持，也不源于无政府状态，如结构现实主义者所持。它源于生成，且它会在一个生产任何东西的体系中出现，无论其成员如何被构成、无论其上的权威如何被安排。

这一主张同时比现实主义自身的更友好也更激进。它更友好，因为它移除了恶行的归咎：无须设定任何胃口，而在国家之间出现的权力，会像在任何进行生产与交换的各类各方之间那样出现。它更激进，因为它封闭了该学科另一半的那个长存的希望。如果权力是生成的代价，那么制度安排未必能逃脱它，而世界国家，终于到来，会发现它把问题输出到了它自己的内部，而非解决了它。

对平等主义

那个提议废除不对称的纲领不仅仅是困难的。它是自我挫败的，其方式是结构性的而非实践性的。在它成功移除差异的程度上，它移除生成所在于其中的那种差异化，因而移除了那个其分布本是此番作为之目的的价值的生产。一个完美拉平的体系是一个几乎不生产什么的体系。这并不谴责平等主义的冲动，本文分享这一冲动。它谴责对该冲动之目标的一种构想，并迫使该冲动去另寻一个目标，而下一节尝试供给之。

对关系理论

世界政治的诸关系理论建立于一个本文分享的前提之上，即关系先于实体、行动者是关系中的行动者而非随后进入交往的诸实质。秦的关系理论是其中发展得最充分的，且它把 阴 与 阳 的关联关系作为其形而上学内核，持其基本状态为和谐，并作结论说关系根本上是和谐的、且能够被作为和谐来管理。该主张奠基于一个对世界之关系性性质的悠久沉思传统，且它不是轻率提出的。

对关系理论的通行反驳，秦已直接应对，追问一种从关系而非从对立的实质出发的理论如何能够辨认主体与客体，以及它随后如何能就二者之间的支配与从属给出说法。秦的回答是消解这一二分。支配者与被支配者之间不可化约之对立的表象，在他的说法中是一副二元论透镜的产物，而 中庸 辩证法，它从一开始就不假定任何二元，通过拒绝张力被提出于其中的那些术语而消解了它。这一步是融贯的，且它是可得的。

本文接受同一关系性前提而返回相反的回答。

支配未必是透镜的产物。它可从该前提推导。进行生成的关系必定要差异化；它们所差异化出来的东西包含供给关系据以进行之条款的能力；而一段其条款由其中一方供给的关系，是一段带有方向的关系。这一推导不需要任何二元论，也不设定任何对立的实质。它只要求关系是生成性的，而这正是一种关于活生生世界的关系理论无论如何都必须维持的东西。

关系本体论并不递交和谐。它递交权力的必然性，并随之递交治理它的必然性。

和谐，在秦所发展的关系理论中，位于根基处。在此处所发展的理论中，它至多位于末端，作为一项成就，而工作的大部分在于详细说明该成就得以可能的诸条件。分歧不在于关系是否居先，两个理论在这一点上一致，且它不是一个关于和谐之价值的分歧，两者都把和谐置于视野之内。它关乎一种关系本体论一旦关系被给予居先地位便有权得出什么结论，而本文的回答是：和谐属于这样一种本体论必须去挣得、而非它可以假定的东西之列。

所欠他人之债

推导的材料并不新鲜，而假装并非如此是徒劳的。

生成生产差异、且差异先于身份而非从身份派生，这属于德勒兹。个体化是一个过程而非一个状态、且个体是一个体系的一个相位而非它的原子，这属于西蒙东。权力是被施用而非被占有、它既是生产性的又是压制性的、且它分布于一个场域之中而非集中于一个顶点，这属于福柯。一个活生生的系统通过与一个它并不指挥的介质的结构耦合而维持自身，这属于马图拉纳与瓦雷拉，以及其后的卢曼。

此处所主张的增量不是本体论。而是那个接榫。

诸块料落于不同的手中。埃格拉什供给规范性主张，即价值理应回到那些生成它的人那里，而没有机会追问不对称究竟为何应当出现，因为在他所研究的领域中它就是给定的。德勒兹与福柯供给差异与涌现的本体论，而不提供任何应当，这在他们身上不是失败，因为他们不是在写规范性理论。秦供给关系性前提并把和谐作为其基本状态。此处所尝试者是那个接榫，而该接榫改变了生成性正义之所是。它不再是可得诸偏好之间的一种道德偏好，而成为对一个抵抗移除的结构性事实的回应。

这所造成的差别值得精确陈述。倘若不对称是偶然的，一处有待清除的败坏，那么对它的正确回应就会是它的消除，而生成性正义就会是”在此期间做什么”的诸提议之一。如果不对称是生成的一个伴随物，那么它可以被塑造而未必能被消除，而它所采取之形态的问题就不再是诸问题之一。它成为那个问题。

反驳

说这是泛权力论。 有人可能说，如果一段生成性关系生产权力，那么这个概念已被膨胀到区分不了任何东西。

回应承认前提而否认结论。权力的存在在生成性系统中是弥漫的，且在存在的层面上这个概念做工甚少。这一点被毫无勉强地承认。区分性的力量在于形式而非存在，而下一节供给据以将诸形式分辨开来的准则。一种只说权力无所不在的理论说得甚少。一种说权力是弥漫的、并且这里就是如何区分建造的权力与消耗的权力的理论，说出了需要说出者，且它需要第一个从句才能说出第二个。

说该推导仅仅是言辞上的。 有人可能说，该链条是分析的，而分析的链条发现不了任何东西。

回应承认论证的性质而争执从中得出的推论。链条是分析的。它的价值在于移除一个原初项而非发现一个事实，而一个原初项的移除改变了可以被追问的东西。一种把权力当作给定的理论，很难去追问它是何种权力，因为它对”权力是何物之一种”没有任何交代。一种从生成推导出权力的理论，可以追问一次给定的权力是否服务于它由之而来的那个生成，而那个问题组织了随后的一切。

说共同议定是寻常的情形。 有人可能说，协调的条款通常是被协商而非被强加的，而条约、公约与常设法律体是证据。

回应是接受该证据而争执它所显示的东西。条款被协商确立了诸方参与了供给它们。它并未确立它们对称地参与了。一次协商在一种语言之内、在一个先例之上、在一个论坛之中、在一张其形状在任何人坐下之前就已定好的桌旁进行，而这每一件都是由某人供给的。问题不是较弱一方是否说了话。而是它被迫据以说话的那些条款是否是它自己造就的。因此，共同议定不是对该说法的驳斥，而是有待解释之物的一个实例，而该说法预测的正是实际所观察到的东西：协商出的结果以尚可容忍的忠实度，追踪供给协商据以进行之条款的能力的分布。

说国家不是生成性过程。 有人可能说，该推导适用于生产、学习与差异化的过程，而国家是否是这样的过程是一个悬而未决的问题。

回应是，这个问题虽是经验性的，却不是一个困难的问题。国家生产技术、知识、制度、文化与能力；它们学习；它们内部地并彼此地差异化；而如此生产出的诸差异，正是进入构成国际政治之协调的那些差异。推导对一个国家所要求者，是它进行生成，而国家进行生成。

§4 两类不对称

如果不对称抵抗移除，那么要紧的问题与其说是有多少不对称，不如说是它是何类不对称，而一种只有一个不对称范畴的理论很难去追问。现实主义有一个范畴。不对称在其中作为能力的差别出现，而能力的差别被读作威胁，因为该理论几乎不提供别的东西供它被读作。本节供给第二个范畴，并随之供给一条据以分辨二者的准则。

区分

一种不对称是生成性的，当诸方有差异、当该差异是互补的、且当关系提升每一方生成的能力时。

一种不对称是攫取性的，当诸方有差异、且当其中一方运用该差异以降低另一方为自身生成的能力时。

二者在标量的层面上难以区分。两者都是不对称。两者都可以呈现同一的能力比率、同一的资源分布、同一的跨关系的货物流动。几乎没有什么能从量值上读出的东西能把它们分辨开来，而一种只读取量值的理论因此会系统地误分类，把第一种当作有待纠正的危险，并把它自身繁荣的条件纠正掉。

准则

一种不对称是生成性的还是攫取性的，取决于跨越一个闭合的互动回路，每一方独立生成的能力的净变化。

准则的分量落在最后一个词上，而这个词必须被给以其力量。

攫取的标志不是较弱一方生产得更少。而是较弱一方生产得更多，且在较强一方所供给的条件下生产。产出上升；独立性下降。关系被体验为有益，被辩护为有益，且在一份计算”已被生产之物”而非”若关系终止则本可被生产之物”的账簿上显示为有益。依赖，在其发达的形式中，是通过供给而非通过剥夺被制造出来的，这正是它为何如此鲜被它所被施用于其上者认出、又如此轻易地被施用它者辩护。

典型的对照是生产线的转移与设计能力的转移之间的对照。第一种提升产出而降低独立性，因为被安装之物能够被操作却较难被修改、扩展或替换，且其操作只要供给方继续供给便持续。第二种提升独立性，因为被传递之物是造出此物的能力而非此物本身，而该能力一旦被传递，便不易被给出它的一方收回。这两种转移可以有同一的货币价值，且在任何金钱所能保有的账目中可以看起来一样。

准则是可观察的

准则是行为性的，且不要求对价值流动的追踪，那在任何情况下都是不切实际的。

人们观察依赖的程度：接受方的生产性活动有多大比例需要对方所控制的投入、许可、标准或授权。人们观察断供之下的存续：当供给被撤回时接受方能继续做什么，且其能力多快地退化。人们观察随后独立创新的速率：接受方在接受之后，是否继续为自身造出东西。

而人们观察一个进一步的检验，即那个区分性的检验。

一个回路是生成性的，如果它在接受方中生产出捐赠方并未设计的诸能力。

这个检验之所以锋利，是因为慷慨并不满足它。一个捐赠方可以给出许多，可以自由地给出，且可以带着一切有益的意图给出，而接受方仍然可能只能做捐赠方装备它去做的事。在接受方开始生产捐赠方并未预料、也无法指明之物之处，某种不在转移之中的东西已在它们之间传递，而正是那次传递，是准则意在捕捉者。

三类回路

准则产出三种而非两种情形，而第三种在现有规范性文献中缺席。

攫取性回路。 至少一方的独立生成能力跨越回路而下降。这是生成性正义被设计来识别的情形，而它正确地识别了它。

维系性回路。 双方的独立生成能力都不跨越回路而改变。取走甚少，建造亦甚少。关系持续，账目平衡，且无人受屈。

生成性回路。 双方的独立生成能力都跨越回路而上升。关系生产出一份剩余，为双方先前皆未持有、且双方皆未持之以对抗对方者。

中间那种情形不在生成性正义据以建立的诸领域中出现，而理由是发人深省的。编发店与合作社在构造上就是生成性的：它们的存在是为了让某物被造出来，而一段生产不出任何东西的造物者之间的关系，会在能被研究之前就已解体。在关系的生成性性质可以被视为理所当然之处，一条侦测攫取的准则是足够的，因为攫取于是就是关系可能失败的唯一方式。

国际关系不可以被以这种方式视为理所当然。一段国家之间的关系可以持续一个世纪而几乎不生产什么，被惯性、被退出的代价、或被双方对一段关系之表象的兴趣所维系。这样一段关系在形式上是干净的而在实质上是惰性的。价值并未从那些生成它的人那里被攫取，且回路已闭合；但几乎没有什么被生成，而回路的闭合是一个几乎没有条目被记入其中的账目的结清。盗窃的缺席并不确立一段关系的状态，而一个被带到”惰性关系寻常可见”之地带的框架，需要第三个范畴，而在它的本土地带它没有机会需要之。

一处不相容

由此有一个关于现实主义的、可推导因而不带论战色彩地陈述的结论。

现实主义的成功准则，即体系不坍缩且能力被对称地分布，与一个生成性回路的标志相张，后者是逐点的不对称在时间上被维持。

一个完全达成现实主义目标的体系，按此处所提出的准则，未必是生成性的。

不作任何”现实主义者渴望压制生成”的主张，也无须任何这样的主张。要点是，一个理论的成功条件趋近另一个理论的失败条件，且这一张力从诸定义而非从任何人的动机推出。

这一实例是现实主义满怀信心地主张为己有的一个。两个集团在近乎完美的对称之均衡中被维持四十年，一个未坍缩的体系，一种双方持续劳作以保持均等的能力分布：按现实主义的准则，这是一次成功，且它已被如此记录。按此处所提出的准则，它充其量是一个维系性回路，而在其大部分范围上是一个攫取性回路，其中每一方把它的生成能力耗费在对方把它锁定于其中的那单一领域上，而每一方生成许多别的东西的能力，只要锁定持续便下降。体系未坍缩，这是真的，且它属于关于它所能说出的最无趣的东西之列。

局限

准则有三处弱点，且它们在此被点名而非被掩埋。

测量。 独立生成能力在原则上是可观察的而在实践中是困难的。上文提出的代理量是代理量，且其每一个都可以被一个知道自己正被它们测量的一方所操弄。此处不提供任何完全令人满意的操作化。

回路的边界。 准则被定义为跨越一个闭合回路，而一个回路在何处闭合的确定本身就是一个判断。在许多情形中它是一个有争议的判断，而那个有权划定回路边界的一方，处于一个与那个有权设定比率的位置并非无关的位置。这一点此处未获解决。

汇总。 三者中最严重的是汇总。一个国家的生成能力是对一个具有极大异质性的内部结构的汇总，且它可以在汇总上上升而对一个地区、一个部门或一个阶级则下降。一次提升某国家生成能力的转移，可以通过把该能力集中于一个自身相对于其自己的边缘是攫取性的都市中心而做到。

因此准则，在国家层面应用，可能把一个在低一层隐藏着攫取的回路证明为生成性的。一个止步于此的框架会在它推却审视的那一层面上冒同谋的风险。所需要的是一种多层次的处理，而本文不提供之。

§5 异质性问题

论证现在抵达它主要关切的那项操作。前几节已从生成推导出不对称，表明它抵抗那种在移除它之后仍让生成存续的清除，并供给了一条据以把建造的不对称与消耗的不对称分辨开来的准则。准则追问，跨越关系的一个闭合回路，每一方独立生成的能力上升还是下降了。它预设回路能够被闭合，且这一闭合能够被查验。随后是一个关于那一预设的困难，且它是一个在国际政治的地带上出现、而不在准则由之而来的那个地带上出现的困难。

国际回路的货载

在国家之间承载价值的关系，跨越诸记法承载它。一个国家供给技术而接受承认。一个国家供给资本而接受正当性。一个国家供给安全保障而接受基地权与战略对齐。一个国家供给一份道歉而接受，或未能接受，一桩历史宿怨的了结。一个国家供给市场准入而接受一个多边机构中的选票。这些之中每一个都是国际生活的一次寻常交易，而在它们每一个之中，出去之物与回来之物属于不同的类。

它们，在相关的意义上，是不可通约的。没有明显的实质，是技术与承认二者的量，也没有度量站在一份安全保障与一项基地权之背后并把每一个表达为如许多的一种共同材料。主张与其说关乎测量的当前状态，仿佛一个足够精细的仪器最终会供给那缺失的尺度。它是一个关于价值本身的主张。它们不是从一个共同来源生产出来的，它们不是作为某一件东西的不同量进入世界的，而汇总未必能恢复一个本来就不在那里的统一。

生成性正义理论由之而发展的诸案例通常不具这一结构。在作坊中，在合作社中，在造物者的共同体中，劳动者所创造之物回到劳动者那里，且它作为收入、作为技艺、作为所有权、作为时间回来。回报可以是足够或不足、及时或延迟，但它是一种与贡献同类的回报，且其充分性因此是一个能被查验的事项。人们计算出去了什么，人们计算回来了什么，而账目在同一货币中，因为它们是关于同一件东西的账目。这不是该理论的一处缺陷。它是该理论被工作出来于其中的那个领域的一个属性，且它是国际领域所不共享的一个属性。

作为判断的闭合

后果是，一个在同质情形中是经验性的问题，在异质情形中成为别的东西。

在回报同质之处，一个回路是否已闭合的问题是一个事实问题。由劳动生成的价值是否回到了劳动者那里？人们查验，并发现它回来了或没有回来。发现可能是困难的，记录可能是糟糕的，因果链可能是漫长的，但困难属于一种更好的查验在原则上能够克服的类型。问题有一个答案，而答案不取决于是谁在问。

在回报异质之处，情况不再如此。为使回路闭合，某物必须被迫算作对另一种物的偿付。承认必须算作对技术的偿付。一项基地权必须算作对一份保障的偿付。承载分量的那个词是算作，而它所命名者是一项操作，不是一次发现。技术之中甚少、承认之中甚少的东西决定了后者的多少了结前者所招致的义务。二者不处于任何比率之中，因为一个比率需要一个据以采取它的共享维度。

在异质情形中，”一个回路已闭合”这一陈述不是对一个事实的报告。它是一个判断，而一个判断需要它据以作出的条款。那些条款并非由被判断的价值供给。

这就是随后通篇将被称为转换的那项操作，而它据以被施行的条款将被称为转换比率。这套语汇是有意平实的。所争议者是那项日常操作，凭借它一个含有不同类之条目的账目被宣布结清，而问题是那一宣布之条款从何而来。

三条回避之路

论证要求转换难以避免。倘若有一种在异质情形中回答闭合问题而不设定一个比率的方式，那么该比率就会是诸方法之一的一个产物，且几乎不承载它自己的分量。三种这样的方式呈现出来，而每一种都必须被封闭。它们按其初始可信度的顺序被列出。

第一条回避之路，货币化

第一个提议是，异质性只是表面的，因为一切都能被定价。技术在金钱中有一个价值，承认在金钱中有一个价值，而一旦每一个都被指派了它的价格，账目就能以寻常的方式被查验。看似阻碍审计的那种异质性，原来是一种仅在表象上的异质性，被那个现代经济体无论如何早已应用于几乎一切别的东西之上的万能溶剂所消解。

有三个回应，且它们是累积的。

第一个是，国际政治主要关切的那些善抵抗定价，且它们的抵抗并非它们之所是的偶然。承认、尊严、主权与历史正义是政治性的善，而它们之为政治性，其程度就在于它们抵抗定价。一个接受它的承认之价格的一方，在接受之际，已让渡了承认本应确保之物的大部分，因为承认所确保者是作为一方而非作为一个数量的地位，而一个可以被以一个价格买断的一方是一个数量。价格的拒绝，与其说是面对一次在原则上本可作出的估价的固执，不如说是被理解为对所争议之善的构成。埃斯佩兰与史蒂文斯，其关于这些事项的说法在下文详述，给出了这一观察的一般形式：存在着一些价值，其不可通约性不是不完善测量的一个残余，而是持有它们者之身份的一个条件。

第二个回应更具破坏性，因为它表明这条回避之路是被回避之物的一个实例。给承认定价，就是确立一个承认据以与金钱结算的比率。这一定价与其说是转换的一个替代品，不如说是它的一个实例。它就是一次转换，在某个比率上施行，且它与其他转换的不同之处主要在于，那个一切都被译入其中的记法是一个货币性的记法，且它是如此熟悉以至于翻译不被注意地过去了。因此该提议不回答比率从何而来的问题。它以一种特定的方式即在金钱中固定比率来回答它，而随后以”这是自然的”为由推却对这一固定的辩护。这是极大量结算的奠基性一步，而它被作出时的那种轻易，恰恰是它必须被拒绝的原因。

第三个回应是经验性的，且值得陈述，因为前两个是概念性的，而读者可能怀疑概念性的顾虑是以事实为代价买来的。如果货币化闭合了回路，那么对政治性债务的货币化结算就会结清它们。它们通常不结清。赔偿被提出并被拒绝。补偿被接受而宿怨仍存，且以一种接受补偿并未触及、有时反而加剧的形式存留。以金钱换取一桩错误的提议，经常被体验为第二桩错误，而这一体验更宜被读作感知而非混淆：金钱据以与那桩错误结算的比率，是由做出那桩错误的一方设定的。一种无法容纳这一点的理论会漏掉它本应解释的东西的一大部分。

第二条回避之路，范畴分离

第二个提议是，转换可以被拒绝。让每一记法按其自身的条款被审计。让技术与技术结算、承认与承认结算，并让任何账目都不被允许跨越它们之间的那条界线。如果诸记法从不相遇，那么就从不需要一个比率。

回应分三部分，而第一部分只是：世界并非如此安排。国际政治的诸回路凭构成而非凭败坏就是跨记法的。发展援助被换取多边机构中的选票。安全保障被换取基地与对齐。市场准入被换取政治上的迁就。这些与其说是一种更纯粹的实践被污染了的退化情形，不如说是实践本身。一种以排除它们为开端的国际价值循环理论，与其说是把它的对象理想化了，不如说是抛弃了它。

回应的第二部分是一个归谬。假设分离被强制推行，且每一记法被单独审计。那么在几乎每段双边关系中几乎每一记法都会被发现不平衡，因为平衡项落在另一记法中。技术流向一个方向而不以技术偿付。承认流向另一个方向而不以承认偿付。在所提议的计账上，国际政治中的大多数关系会在大多数记法中是攫取性的，而结论是荒谬的。荒谬性起诉的是计账而非诸方，且它在计账拒绝转换的那一点上起诉它。

第三部分最短。分离不回答结算问题。它禁止这个问题。说诸记法不可彼此结算，与其说是找到了一种结算它们的方式，不如说是宣布结算不发生，在一个结算持续发生、且构成所研究之关系之实质的世界中。一条通过否认它的待释项而成功的回避之路，更宜被描述为一次拒绝。

第三条回避之路，主观满意

第三个提议是三者中最谦逊的，且正因如此是最诱人的。让诸方决定。回路已闭合，当双方都同意它已闭合时。无须一个外在的比率，因为比率就是诸方在事件中所接受者，而一项双方都接受的结算凭这一事实就是一项结算。

回应是，这条准则把它本应评价的那件东西本身，当作它的证据来征询。

考虑什么决定了一方是否接受一项结算。它接受，当它所接受之物，在它自己的估价图式之内，登记为对它所给出之物的充分。但它的估价图式并不独立于关系。它是在关系之内、随时间、在这样的条件下形成的：充分性的条款被较强一方持续排演、被较弱一方持续栖居。因此较弱一方的满意是比率的一个效果，而以它作为比率的检验，就是以比率自己的后果来确证比率。这一循环不是一个微妙的循环，而措辞的精细未必能避免它，因为它在于境况的结构而非检验的措辞。

殖民案例是范例，且它是为何诉诸真诚未必能拯救这条准则的原因。对结算满意的被殖民主体，并不因此被表明已接受了一项正义的结算。满意能够被生产出来，属于一段漫长关系的成就之列，而一种以满意为其最高法庭的正义理论，在决定性的时刻几乎无话可说。它所拥有者是一份关于关系在生产那个确证它的心性方面之成功的报告。

应当补充的是，这条回避之路如果被接受，会凭构造使后面某一节所审视的机制不可见。不利比率如何被接受这一问题，是本文视为核心的问题。一条把接受当作探究之终点的准则，已事先决定了这个问题将不被追问。

诸命题

三条回避之路是那些呈现出来者，而每一条都已失败。货币化施行了它宣称要避免的那项转换。分离禁止了它宣称要回答的那个问题。满意以比率的效果作为比率的检验。所剩下者可以被直接陈述。

命题 5.1（转换的必要性）。 在异质记法的价值循环于一个关系体系中之处，”一个回路已闭合”这一判断预设了那些记法之间的一个转换比率。

命题 5.2（比率的被设定性）。 转换比率不可从它所转换的价值推导，因为异质记法的价值彼此不处于任何比率之中。因此比率是被设定而非被发现的。

从这两条推出一个将统辖本文其余部分、且使它付出甚多的推论。

推论 5.3（无正确的比率）。 不存在正确的转换比率。一种预设了一个的批判，已引入了一种对所争议价值不可得的共同尺度。

代价应被立即陈述，且它会在到期时被再度陈述。推论 5.3 从本文手中移除了剥削的古典词汇。那套词汇通过与一个真实比率相比较而运作。劳动价值论供给这样一个比率，而它正是允许不平等交换理论说”一次以某种价格比率进行的交换是不平等的”的东西，因为不平等是相对于一种度量的关系，并要求一种度量存在。一种否认该度量的批判无从说比率是错的。它已让渡了一个比率可能是错的那种意义。

本文有意接受这一让渡，理由是替代方案更糟。一种保留共同尺度的国际价值理论，以构成国际政治的那些价值为代价来保留它，而它们不是由劳动生产的，也不是任何实质的量。一种能说咖啡被定价过低、却对一份保障是否被偿付、一份道歉是否被欠下、或承认是否被过于廉价地购买完全无话可说的理论，已通过把它的对象窄化到该对象不再包含它本应解释之物，而买来了它的严谨。度量的让渡是言说国际政治所关乎之事的代价，而度量既去之后批判还剩下什么，正是本文规范性一节所致力的问题。

论证的界限

论证的界限应被平白陈述，因为本节承载着本文的分量，而此处的一次夸大会使它坍缩。

尚未被表明转换是不正当的。转换发生，它们难以避免，而此处所说的甚少东西不利于它们。尚未被表明一个特定的比率是错的，且按推论 5.3 没有任何这样的东西能被表明，因为不存在一个比率能相对于它被发现有所欠缺。

已被表明者更狭窄且足够。在一次跨记法的结算中一个比率是在场的；它是被设定的；且它不可从它所结算之物推导。由此推出，在这样一次结算中有人做了那次设定。谁做的，以及占据那个据以做设定的位置意味着什么，是随后诸节的主题。

§6 拒绝被封闭之处的结算

前一节确立每一次跨记法的结算都预设一个被设定的比率，并以”某人必定做了那次设定”这一观察作结。人们会禁不住直接进而命名一个权力，而这一诱惑应被抵抗，因为该权力已被命名。因此本节以陈述它所继承者为开端，而它继承了甚多。

继承

本文称之为转换的那项操作，就是社会学称之为测度化的那项操作，且它以那个名字已被研究了将近三十年。埃斯佩兰与史蒂文斯把测度化定义为，把通常由不同单位所表征的诸特征依据一种共同度量来表达或测量，并观察到它把质转化为量、把差异转化为量值。他们把说法的一节明确献给测度化如何折射权力关系的问题，而他们回答：它是权力的一种样式，其权力既是策略性的又是构成性的，它重构权威关系、创造新的政治实体、确立新的解释框架，且它决然不是其倡导者所主张的中立的技术过程。他们观察到它的运作通过例行化而变得不可见，以至于它所需要的劳作以及围绕它的那些假定被遗忘，且测度化构成它所自称测量之物。他们把马克思读作持”劳动是伟大的测度者、且价值通过抽掉特定工作诸形式的独特性而进行测度”的观点。而他们观察到，在一段直接关涉当前主题的段落中，被理解为妥协之艺术的政治本身就是测度化的一个宽泛实例，因而利益集团的讨价还价预设了那项它无从审视其条款的操作本身。

本文不重新推导上述内容。它站在其上。设定结算之条款乃是权力的一种样式、该权力是构成性的而非仅仅是策略性的、且其运作被它的例行性所遮蔽，这一主张属于测度化社会学，并被毫无保留地归功于它。

本文所补充者落在别处，而它可以通过留意社会学文献所行进的方向而被精确定位。

边界与内部

测度化社会学据以建立的诸案例是侵入的案例。一种尺度挺进一个曾被隔离于其外的领域，而由之出现的政治问题是这一挺进是否将被允许。经济学家尝试为一项曾未被定价的活动合成一个价格。可比价值的倡导者尝试把家务劳动纳入工资的尺度之内。法学者尝试确立，或抵抗，一个儿童市场。在每一种情形中，一种估价样式正压向一个另一种样式曾占上风的地区，而戏剧是一场边界的戏剧。

抵抗，在这样的情形中，往往采取一种标准形式。它是不可通约性的主张。追随拉兹，埃斯佩兰与史蒂文斯把这一主张的一个强版本命名为构成性不可通约者：一个其不可通约性与持有它者之身份缠结在一起的价值范畴，以至于允许那次权衡取舍就是变成另一个人。亚瓦帕伊人，其祖传土地曾受一座大坝威胁，把那片土地作为一个构成性不可通约者持有，而官僚据以评价大坝的模型要求那片土地被弄得与决策的其他成分可通约。这一要求不仅仅是不便的。它是对所争议价值以及那些价值所界定之身份的一次矛盾。拒绝把他们的孩子想成可替代者的父母、拒绝让他们的生活相对于职业成就被定价的持家者、以及拒绝把他们的承诺带上谈判桌的运动，都在作同一步，以不同的规模、带不同程度的成功。他们在说：这不是可以拿来与那个交换的那类事物。

埃斯佩兰与史蒂文斯甚至点出那个外交实例，且他们恰恰以这种方式读它。以色列与巴勒斯坦两族人民的领土承诺，他们观察到，已在外交谈判桌上挫败了无数次测度领土利益的努力。该案例在他们的说法中作为一个拒绝的案例出现，而拒绝正是使它对他们有趣的东西。

测度化社会学是一种关于边界的理论。它的主题是一种尺度的挺进以及抵抗它的那次拒绝，而它对抵抗的说法所依赖的那个政治可能性，是说出这一话语的可能性：不是这个，不在这里，不按那些条款。

国际政治，在其寻常运作的大部分中，发生于内部。边界很久以前就被跨越了，且它不易被重新跨回。诸方已在被结算，持续地，而摆在它们面前的问题不是是否允许该尺度，而是该尺度将按谁的条款运行。

拒绝的被封闭

本节的主张现在可以被陈述。社会学文献所记录的那次拒绝需要一个据以能够被维持的位置，而一个缺乏从关系中退出之途径的国家很难占据它。

考虑拒绝在社会学案例中成就了什么。亚瓦帕伊人，在宣布他们的土地不可通约时，并不因此阻止大坝。他们可能失败。宣布所确保者，无论它是否胜出，是那次损失作为一次损失而非作为一次结算被登记。他们并未因那片土地而得到偿付，因为他们已拒绝了偿付本会是偿付所据以进行的条款。错误仍是一桩错误，且仍可被作为一桩错误来命名，因为本会了结它的那个计账被拒绝了。

一个国家很少处于能够如此的位置。它可以宣布其主权不可通约，而它会不顾此而被结算。它可以主张任何数额的投资都不能了结一次殖民所招致的义务，而账簿仍会被拉出来，投资记于一侧，攫取记于另一侧，且由无须其同意的诸方宣布账目结清。宣布被听见了。账目无论如何被结清了。

原因不是国家比共同体更弱，它们并不更弱。原因是亚瓦帕伊人可得的拒绝是一次不参与的拒绝，而参与是一个国家最不能推却之物。拒绝结算就是从结算得以做成的场域中退出，而从那个场域中退出，就是成为，不是一个已拒绝被结算的国家，而是一个不登记为一方的国家。机制在下一节被论及。此处记录这一结构性事实便足够了。

命题 6.1（拒绝的被封闭）。 在一方若不退出关系便不再算作一方之处，不可通约性的主张作为一种抵抗形式对它是不可得的。测度化于是不再是一条有争议的边界，而成为关系的一个构成性条件。

于是国际的情形并非社会学所描述过的那个现象的又一实例，而是它的极限，在此极限上该现象改变了其性质。在拒绝可得之处，测度化是一条争议中的边界，而政治问题关乎边界的位置。在拒绝被封闭之处，边界不在争议中，而政治问题关乎内部据以被治理的条款。

拒绝的被封闭对后面所发展的规范性论证有一个后果，且值得在此标出，因为它把一个提议转换成一次推导。在抵抗的强形式不可得之处，所剩下者是对比率的争议而非对它的拒绝。下文所提出的规范性条件，与其说是因其吸引力而从诸替代方案中被选出，不如说是在其他方案倒下时留存下来的东西。

针对权力的诸面相

测度化文献尚未把它的对象置于权力的标准分类学之上。这一置放值得施行，既为定位该操作，也为预先阻止”它是一个已知者的一个新描述”这一反驳。所继承的每一种权力形式被依次审视，而结算的操作被发现落在它们每一个之旁而非落在其中任何一个之内。

达尔的强制性权力让一方去做它本不会做的事。比率的设定不强制任何行动。它作用于账目而非作用于意志。一方可以自始至终自由地行动，选择它自己的每一步棋，并在末了发现它的账目已在一个它并未设定的比率上被结算。

权力的第二面相控制议程并把诸事项排除于其外。此处没有什么被排除。一桩宿怨可以被提起、被列入、并被记入账目。比率所决定者与其说是什么可以被说，不如说是一件已被说出之物相对于另一件东西值多少。一桩宿怨可以被充分地道出、充分地被承认、并充分地被记入，而在一个被冤者并未设定的比率上被结算。

卢克斯的第三面相塑造欲求，以至于宿怨不出现。此处所审视的操作作用于诸等价而非作用于欲望，而在寻常情形中二者分开。一方可以欲求它在未被扭曲之形成条件下本会欲求者，可以清晰地表达该欲求，可以精力充沛地推动它，并可以发现它已被一份不同类的提议所了结，因为那一类据以与这一类结算的条款不是它可设定的。第三面相阻止那个抱怨。比率给它定价。

基欧汉与奈在此处接近了，而这一点值得小心。他们的洞见是，一次依赖的不对称本身就是一种权力资源，且这种权力是可与军事能力相分离的。但他们所描述的权力是一个标量：一个谁能更好地容忍关系断裂的问题，沿那个容忍的单一维度被测量。讨价还价的筹码预设争议中的诸事项已被弄得足够可比以至于可以就其讨价还价，而比率的设定正是把它们弄得如此可比者。因此比率先于讨价还价而非它的一个种类，而那个把问题化约为筹码的框架，已施行了它自己的一次转换而不加评说。

葛兰西式霸权解释一种秩序如何被同意。此处所争议的操作是那一项，凭借它一种善成为对另一种善的结算。正当化是那项操作的一个后果，而霸权理论描述后果，同时把操作留作未供。

布迪厄比其余各面相更近，且需要小心处理。他的符号权力把一种任意的秩序自然化，并通过误识而确保其被接受。但误识是一种相对于一个已就位之安排的关系，而自然化预设有某种东西要被自然化。布迪厄极有力地解释了比率何以不被质问。他把”何以竟有一个比率”留作未解释，而这两个问题是不同的。

奈的软权力是一份吸引力的存量，一份被一方持有、并以寻常方式耗费的资源，因此它属于本文一直着力要抛在身后的那个标量传统。此处所争议者与其说是一份资源，不如说是一个位置：那个据以宣布诸等价的位置。

账簿呈正数的殖民地

框架在对手框架沉默之处显示其用处，而有一个案例，它们在其中沉默。

殖民统治的核心操作不是劫掠。劫掠是可见的、可计算的、可抗议的，且它经常被争执、偶尔被遏制。核心操作是比率的单方设定，而它被保存在殖民行政为自身所拟的那句话里：我们给了文明、铁路、法律与医药；你们回报了资源、劳动与土地；账目结清。

承载论证的那个词是结清。它预设了一个铁路据以与主权结算、医药据以与土地结算的比率。比率是由宗主国设定的，而其公平性本可被争执所据以进行的语言也是由宗主国设定的，以至于希望反对的被殖民一方被迫以”一条铁路值多少”的术语来进行其反对，而那正是那个已被回答的问题。

现在考虑物质账簿呈正数的情形。存在着一些殖民关系，其中净资本向内流动，其中铁路被建造、学校被创办、死亡率被降低，且其中一次诚实的物质流动计账表明该殖民地接受的多于它送出的。

对手框架很难找到那处不义。

生成性正义找不到它，因为价值流向生成者而非流离他们，且回路，在物质上，闭合了。不平等交换理论找不到它，因为它通过一个负余额侦测剥削，并需要劳动价值论来确立该余额是负的；在余额是正的情况下该理论几乎无从下手。而测度化社会学看见运作中的权力，却按其自身的方法论承诺而推却称它为不义：它观察到，带着精心研究过的不偏不倚，测度化可以扩大决策，也可以为一个事先已定的决策提供正当性，它可以被精英操纵，也可以约束他们，且它止步于此。

这个框架找到那处不义并定位它。它既不在于流向，流向反了过来，也不在于比率的不公平，因为按推论 5.3 不存在一个公平的比率能相对于它被发现有所欠缺。

它在于任何可以从中重开该比率之位置的被封闭。

作为结算体制的现实主义

在论证进一步展开之前还需要一个应用，而它关乎该论证被提出于其中的这门学科。

现实主义把承认、正义、身份、文化与正当性转换成单一的标量，它称之为能力，且它主张这些东西之关涉于国家的行为，其程度、且仅其程度，在于它们能被如此转换。这是一次测度化，且它在某个比率上被施行。

现实主义之为一门科学的主张，正建立在其不承认这一点之上。它把该转换呈现为对何者驱动国家的描述：到头来，它说，只有权力算数。但一切别的东西能够被结算于权力之上，与其说是一项关于世界的观察，不如说是一项关于账目条款的规定，且它是该领域的奠基性规定。转换先于诸发现，而诸发现不检验它，因为它们被记录在转换所确立的那个账簿中。

有一个后果，且它解释了某种被长久注意到却从未被交代过的东西。当现实主义，在其规定性样式中，告知一个国家其承认的主张几乎不算数除非有能力支撑，它与其说是在报告一条规律，不如说是在强制推行一个比率。现实主义规定倾向于带来它所描述的诸条件，这一点自该学说存在以来一直被评说，且一般被归因于政治家的心理或安全困境的动力学。此处所提出的机制更简单。一种供给账目据以被结算之条款、并被那些结算账目者所相信的理论，无须说服任何人任何东西便能造就它所描述的那个世界。它只须被使用。

§7 比率之据

本节以低于前几节的置信度被提出。它推进一个机制，且它推进之，是因为有一个待释项是可得的机制所不解释的，而不是因为它已被证明。论证的地位在本节末尾被陈述，且应在通篇被铭记。

被留置未决的问题

如果比率是被设定而非被发现的，且它因此在它所结算之物的本性中没有任何权威，那么它为何不被简单地争议成对称？两方隔着一个其条款是任意的账目彼此相对。每一方都有一个对有利于自身之条款的利益。为何其中一方的条款占上风？

可得的答案是不充分的，而每一个都因一个发人深省的理由而有所欠缺。

强制是不充分的，因为比率在强制已消退之处存续。后殖民国家的形式主权是完整的，而它的账目继续在别处所确立的条款上被结算。倘若比率被武力固定住，那么它在武力撤除之后的存续就会需要解释，而几乎没有解释可得。

讨价还价的权力是不充分的，因为它预设了那个有待解释者。就一项结算讨价还价，就是在一套已确立的等价之内讨价还价，而问题是等价从何而来。一个把比率奠基于讨价还价权力的说法，已把结果置于原因之前。

信息是不充分的，因为比率在充分信息的条件下存续。无人被欺骗。较弱一方完全能够陈述条款是任意的、它们是由较强一方设定的、且它对它们的接受服务于较强一方的利益。它陈述这一点，常常带着极大的精确，并接受那些条款。

判断的所在

提议从该操作的一个至此未被审视的特征出发。一个等价与其说是它所关联的两件东西的一个属性，不如说是一个判断，而一个判断是从某处作出的。

说一项基地权算作对一份安全保障的偿付，与其说是报告一个关于基地权与保障的事实，不如说是说，在某个计账图式之内，一侧的条目了结另一侧的条目。图式在某处。它被占据，而占据它的那一方是那个”事物为之而算数”的一方。

命题 7.1。 设定转换比率的权力，与占据”事物相对于其而算数”的那个位置在结构上相联系。无论谁占据那个位置，都设定比率，不是通过强制对它的同意，而是通过成为那个”账目为之而是一个账目”的一方。

这一结构已在精神分析的语汇中被小心地工作出来，其中那个”一个主体的诸生产物为之而算数、并从之接受其作为生产物之价值”的位置，是大他者的位置。论证中几乎没有什么东西要求接受这一语汇随身携带的临床装置，而一个推却该装置的读者可以把随后者读作一个关于承认之语用学的主张，且几乎无所损失。

不利比率的被接受

机制通过解释一个现实主义处理得糟糕、且自任何人观看以来一直显而易见的现象而挣得其位置。

现实主义预测一个拥有退出选项、并面对一项不利结算的国家会退出。国家通常不退出。它们留在其条款它们能够表达为不义、其任意性它们能够演证、且从中退出的物质代价本可承受的关系中。这一行为既不是边缘的也不是限于弱者的。它是国际体系的寻常行为，而一种预测相反者的理论，欠一份对该预测何以如此经常失败的说法。

提议是这样的。

从关系中退出就是从一个人为之而算数的场域中退出。

退出就是成为，不是一个已拒绝一项不利结算的国家，而是一个不登记为一方的国家。关系与其说是账目被保有于其处的场址，不如说是”作为一个有账目之物”的条件。

因此较弱一方接受一个不利比率，以便留在计账之内。它以那份剩余购买它的可读性。

这闭合了前一节留置未决的一处空白。拒绝的被封闭曾在那里被记录为一个关于国际体系的结构性事实，而一个不带机制被记录的结构性事实是一个等待解释的观察。机制现在被供给。拒绝被封闭，是因为它花去拒绝者作为一个行动者的地位，而作为一个行动者的地位与其说是结算中所争议诸善中的一个善，不如说是”拥有任何被争议之物”的条件。

由此推出，较弱一方的接受更宜被读作一次购买，而非意志的软弱、误算或虚假意识。它所购买者是那件坚持不接受也不会获得的东西。

一个可检验的推论

提议有一个把它与替代方案区分开来、且可能对它不利的蕴含。

如果对一个不利比率的接受是以地位为交换而买来的，那么接受就应追踪该方为承认为一个行动者而对关系的依赖，且应不追踪、或仅弱地追踪它的物质依赖以及物质替代方案的可得性。两个物质依赖相等、物质替代方案相等的国家，应在其接受不利结算的意愿上依”其作为诸方之地位有多少由所争议的那段关系构成”而分化。

如果相反，接受被物质依赖完全交代，那么此处所提出的机制就是多余的，且应被丢弃。本文中几乎没有别的东西取决于它。

提议的地位

本节提议一个机制。它不证明一个机制。

“那个据以宣布诸等价的位置在结构上是那个事物为之而算数的位置”这一主张，是作为一个具有解释性延伸的假说而被推进的。它的主要支持是，它交代了替代方案所不交代的一个现象：在退出物质上可得、且无人被欺骗之处，对不利比率的接受。那是一个去认真考虑它的理由。它不是一个证明，也不是作为一个证明而被提出。

§8 可争议性

被放弃的据

推论 5.3 否认存在一个正确的转换比率。这一否认是被要求的，因为循环于国际关系中的诸价值不处于任何比率之中，而测量未必能恢复一个本来就不在那里的比率。但该否认移除了批判据以经典地站立的那个据。

针对一次交换的批判的古典形式是，该交换是在错误的比率上进行的。这正是不平等交换理论所说，而它能说它，是因为它拥有一种度量：劳动时间供给那个价格相对于其被发现有所欠缺的真实价值，而这一发现是一次发现，不是一个偏好。把度量拿走，那个句子就失去了它的意义。人们无从说一个比率是不义的，在没有一个比率能相对于它被比较之处，而本文已说没有。

一种已抛弃了它据以批判之标准的批判理论处于严重的困难中，而读者会已经注意到。这一困难是论证最艰难的时刻，且它不被抹平。

本节其余部分尝试一次恢复。它应被作为一次尝试来评判。

抵抗的残余

从测度化社会学已经确立的关于抵抗的东西出发。

在一种尺度挺进一个曾被隔离于其外之领域处，抵抗往往采取一种特有的形式。它是不可通约性的主张：断言这不是可以拿来与那个交换的那类事物。亚瓦帕伊人对他们的土地作出此断言。持家者对他们的生活相对于职业成就的尺度作出此断言。运动在拒绝把他们的承诺带上谈判桌时对他们的承诺作出此断言。这是拒绝的强形式，而在它能够被维持之处它就是恰当的那一形式，因为它并不争执一项计账的条款，而是拒绝进行这项计账。

强形式需要一个据以能够被坚持它的位置。该方必须能够走开并保持其原样。

一个若不退出便不再算数的一方，很难作出那个主张。它会被结算，无论它是否同意，而它”账目不可被拉出”的宣布被听见，而账目被拉出。

可争议性是拒绝已被封闭时抵抗所剩下的东西。

它不是一个为方便而选择的较弱替代品，也不是从一个本可被维持的实质性标准的一次程序主义退却。它是留存下来的东西，而它的必然性是被推导而非被偏好。一个不能说这不是那类事物的一方，仍可以说不在那个比率上，且不单凭你的计量。

条件

命题 8.1（可争议性）。 在没有真实转换比率可得之处，一项结算的规范性条件不是比率的正义，而是它的可争议性：比率被应用于其上的那一方是否拥有地位、语言与论坛可据以争议它。

这三个组成部分各自被要求，而第二个是苛求的那个。

地位是争议比率的被承认的资格，且它要与”就其应用申诉”的资格相区分。一个可以在一个特定案例中恳求宽大、却不可以质问该类案例据以被裁决之条款的一方，已被给予该物的表象而非该物本身。

语言是争议可据以进行、且不完全由设定比率之一方供给的诸术语的可得性。这就是前几节已表明通常被违反的条件，且以那种使违反最难被看见的方式被违反。希望争议铁路据以与主权结算之比率的被殖民一方，必须以”一条铁路值多少”的术语来进行争议，而一条铁路值多少的问题在比率被设定时就已被回答。与其说反对被压制，不如说反对为了被作出，必须以争议中的那枚硬币被作出。

论坛是一个争议被听取、且它能有效果的场址的存在。一个哪里都提不起的争议，已以不可能之物被允许的那种方式被允许。

针对程序主义的指控

会有人反对说，这把正义化约为程序，且一项结算可以是完全可争议的而完全恶劣。

回应不是程序对正义是充分的。而是在没有实质性标准可得之处，可被重开这一属性是一个比率所能拥有的、其本身并非一项设定的唯一属性。一个比率之内容的诸特征是由某人放进去的。而它对被重开的开放性却不是。

一个不能被重开的比率并不因此被表明是错的，因为几乎没有什么东西会表明它是错的。它因此被置于被它所约束的诸方够不到之处。

那，而非它的内容，才是构成不义者。

该提议是对规范性传统的一次实质性修订，且应被作为一次修订来陈述。它把对称从正义的诸条件中移除。一项结算可以是根本上不对称的，按人们乐于构造的任何计账都朝一个方向转移远多于另一个方向，而仍然满足此处所提出的条件，只要它据以达成的条款仍对被它所约束者开放。而一项结算可以是一丝不苟地平衡的而未能满足该条件，如果那一平衡是在一方独自确立、而另一方既无地位、也无语言、更无论坛可据以质问的条款上达成的。

诸后果

三者直接推出。

第一个是，一个带有不可争议之比率的形式上闭合的回路是攫取性的，无论物质账簿显示什么。这就是为账簿呈正数的殖民地所提供之诊断的操作形式，且它正是允许那一诊断根本得以作出者。

第二个是，对称已被从正义的诸条件中移除，如上所述。

第三个关乎制度设计，且在下文进一步发展。要向一个制度提出的问题不是它是否公平地分配。而是嵌入其中的诸转换比率能否被它们所约束的诸方重开。

局限

三处弱点，而第三个是严重的。

阈值。 多少可争议性够了？提议不说。地位、语言与论坛允许有程度，而未给出对每一个多少够了的说法。这一空白是真实的，且它此处未被填补。

俘获。 可以存在一个论坛，其中诸争议被听取并被可靠地输掉，且其中该论坛的存在服务于把损失正当化。条件，如所陈述，被这样一个论坛满足，而它不应被满足。所需者是一个进一步的详细说明，把一个听取的论坛与一个吸收的论坛区分开来，而它未被供给。

它是否强到足以谴责。 假设一个殖民比率曾是可争议的；假设被殖民者曾拥有地位、语言与论坛；假设该比率曾被争议并被维持。那么该结算那时就会是正义的吗？

本文给出的答案是，它那时会是正当的，以一种它当时不是的意义，且该区分是真实的而非言辞上的。已被听取并已输掉的被殖民一方，相对于它的损失，站在一个不同于那个未被允许说话之一方的关系中。但答案是令人不适的，而这一不适应被承认而非被论辩掉。一个发现该框架在这一点上过于昂贵地买来其一致性的读者，并未犯错。他已识别出那个代价。

§9 迈向一种形式化表征

本节是一个研究方向而非一个结果。本文的实质性主张是以自然语言陈述并辩护的，而无一取决于随后者。它被纳入，是因为问题的结构与一套现有的数学装置匹配得异乎寻常地好，而注意到这一点却什么都不说会是奇怪的。

动机

组织了论证的那些操作共享一种性质。不处于任何比率之中的诸价值仍必须循环于一个单一回路中。比较必须在没有一个全局标准的情况下作出，因为对于所争议的诸善没有这样的标准可得。而要紧的比较，与其说是一方与另一方的比较，不如说是一方与它自身的比较，在关系的一次回路之前与之后。

这些正是连络的数学被发展来对付的那些操作。一个连络是那个允许跨越一个缺乏全局框架之空间的诸点作比较的对象。它说一个在一点上被定义的量将如何被携带到另一点，且它这样做而不要求不同点上的量属于一个共同的集合。本文所提出的问题，与另一种语言中的同一问题相近，而这一巧合值得记录。

草图

设基空间为诸行动者之间的关系之图。

设每个行动者之上的纤维为该行动者的生成能力之空间。诸纤维不跨越诸行动者被比较，也不被要求彼此可通约。这与其说是该构造的一处局限，不如说是它的要点。该框架既不要求也不允许我们说一项技术相对于一次承认值多少。

设一个连络由能力沿一条边的输运给出：一次互动，发生于两方之间，改变每一方所能生成之物。

那么一个闭合回路的和乐是一方的能力在被携带绕回路一周并返回它由之出发的那一方之后的净改变。比较是局部的。它把一方与它自身在它自己的纤维中相比较，且不需要任何跨方的度量。

和乐的符号就是第四节的准则。正：每一方现在都能独立生成它先前所不能者。零：维系性回路。负：攫取。

该构造的吸引力是精确的，而它买来什么值得陈述。那个击败不平等交换理论、并迫使本文放弃剥削之古典词汇的不可通约性，在此被结构性地容纳而非被假定掉。一种价值的共同尺度并不被需要，为了追问，就每一方，在它自己的纤维中，它现在是否能生成比它先前所能者更多。

诸困难

四个，且它们并不小。

生成能力的排序。 第四节的准则需要更多与更少。在生成能力之空间未被排序之处，和乐的符号是未定义的，而该构造几乎无所言说。生成能力允许一种排序这一点并不显然，且它此处未被确立。能力可能只是被部分排序的，在那种情况下某些成对的回路会是不可比的，而准则会适用于比已被设想的更少的情形。

连络的详细说明。 说一次互动输运能力，目前是一次重新描述而非一次详细说明。在输运规则被给出之前，该形式化几乎没有内容，而一个没有内容的形式化是一幅图画。

转换的位置。 第四个困难承载最广的后果。如果异质价值的结算是本文的核心对象，那么形式化就必须为它找到一个位置。自然的建议是，转换比率就是那个连络：它是那个”在一个纤维中被给出者据以在另一个纤维中被登记”的规则，且它恰恰是一个连络所是的那类东西，既非源的一个属性也非目的地的一个属性，而是一个据以在它们之间携带的规则。

倘若那一等同被弄得严格，本节就会成为本文的脊柱而非它的附录，而被设定之比率与一个被选择之连络的等同，会供给一种”一个比率是任意的”的形式意义：一个丛不携带任何典范连络，而对一个连络的选择是一次选择。

它此处未被弄得严格，而读者不应被留下一个相反的印象。

装饰的风险。 一个不许可任何在别处不可得之推论的形式化是装饰，而政治理论著作中的装饰比一无所有更糟，因为它以那件东西为代价买来一个严谨的表象。该构造挣得其位置，如果它许可一个”以言语进行的论证所不许可”的结论。

地位

本文不证明该形式化挣得其位置。它陈述那一对应以及诸困难，并把事情留置开放。它是作者进一步工作的主要对象。

§10 对治理的蕴含

上文所推进的规范性条件对制度设计有诸后果，而这些后果不是改革文献通常追求的那些。本节陈述它们，然后陈述那个阻挡在它们之途中、且本文不解决的问题。

设计问题

如果一项结算之为正义，其程度在于其比率仍对被它所约束者开放，那么要向一个制度提出的问题不是它是否公平地分配。

嵌入这个制度中的诸转换比率，能否被它们所约束的诸方重开？

三个观察

第一个是，国际制度特有地把它们的转换比率固定在它们的构成性文件中，且特有地把那些条款弄成它们自身中最不可修订的部分之列。投票权重、否决结构、可受理的诸类主张与不可受理者、一种记法中的贡献据以被记入另一种记法中之义务的条款：这些与其说是程序的规则，不如说是比率，而它们被牢固地嵌入，且比该制度所包含的大多数东西更深地被牢固嵌入。

第二个是，标准的改革辩论关乎在一个固定比率之内的再分配。更多的席位、更多的选票、更多的话语权、一个其分配条款不在争议中的分配的更大份额。在先的问题是谁可以重开该比率，而它鲜被追问，部分因为该比率不把自身呈现为一个比率。它把自身呈现为那个机构的宪制，也就是说，呈现为那个”问题被提出于其中”的东西，而非一个可被追问的东西。

第三个是一条试探性的设计原则，作为这样一条而被提出。

一个制度应把它的诸转换比率置于一个较低的牢固嵌入层级上，而非，如通例，一个较高的层级上。

理由是，分配错误是可见且可纠正的，而比率错误则两者皆非。一个已接受了过小份额的一方能说出来，且以该制度供给的术语。一个在它并未设定之比率上被结算的一方，很难以那些术语说出来，因为那些术语正是它希望争议者。因此最难被看见的条款，正是本应最易被改变的条款，而现存的诸制度把事情弄反了。

退出问题

生成性正义理论在决定性的点上依赖于那些生成者的能动性。他们可以离开。他们可以组成一个合作社，保有他们所造之物，并按他们自己的条款重组回路。该理论的大部分价值系于这一可能性，而在该可能性成立之处该理论是有力的。

在国际体系中，退出被封闭，而第七节的论证提议它以一种比物质意义更深的意义被封闭。一个退出的国家成为，不是一个已拒绝一项不利结算的国家，而是一个不登记为一方的国家，而它由此所失，与其说是关系的收益，不如说是收益被拥有于其中的那个地位。

那么谁来重组回路？

答案未必是诸方自己，而那正是生成性正义所给出的答案。本文没有别的答案可提供。

谁来重组回路的问题是该论证的主要悬而未决的问题，且它不是一个小问题。它关涉生成性正义理论是否经受得住向国际层面的转置，或者它是否仍是一种关于门开着的共同体的理论，并在被带到一个门关着的体系时，成为一份对一个希望的描述而非一个纲领。

层次问题

第二个问题，上文已记录并在此作为治理问题重述。

一个回路可以在国家层面是生成性的而在低一层是攫取性的。一次提升某国家生成能力的转移，可以通过把该能力集中于一个”相对于它自己的边缘一如捐赠方相对于它”的都市中心来提升它。一个把该回路证明为生成性的制度，那时就会证明了一次它未曾审视的攫取。

框架最终必须是一个多层次的框架。它此处不是一个，而在它成为一个之前，它应被带着”它可能在它推却审视的那一层面上表现为同谋”的知识来应用。

§11 作为检验的诸案例

随后的诸案例是检验而非例证。每一个都因它在此框架与一个被点名的对手之间作出区分而被选择，且每一个都被陈述得可能对该框架不利。在一个案例会未能作出区分之处，它被略去，无论多么生动。

生产线与设计能力

这检验第四节的准则。

两次货币价值相等的技术转移被作给可比的接受方。在第一次中接受方获得操作一个过程的能力；在第二次中它获得设计一个过程的能力。预测是，二者会在随后独立创新的速率以及断供之下的存续上分化，且会在物质账簿同一、所转移的金钱相同之处也如此分化。

什么会证伪它。倘若过程转移的接受方显示出与设计转移的接受方相同的随后未预料创新的速率，那么产出与独立能力之间的区分就是在测量甚少的东西，而准则应被抛弃。

账簿呈正数的殖民地

这一次同时对照三个对手检验第六节，且它是本文的主要案例。

该案例之所以被选择，恰恰因为物质计账朝有利于该殖民地的方向进行：净资本向内、基础设施被建造、学校被创办、死亡率被降低。生成性正义应在此处找不到不义，因为价值流向生成者。不平等交换应找不到，因为它通过一个负余额侦测剥削。测度化社会学找到权力，而按其自身的承诺，推却称它为不义。

此处所推进的框架找到那处不义，并把它定位于任何本可从中重开该比率之位置的被封闭。

什么会证伪它。倘若被表明这类情形中的不义在计账被正当地做过之后被一个负流动完全交代，那么比率就是在解释中做工甚少，而框架是多余的。本文关于增量的主张的大部分系于这一案例，而该案例是带着这一知识被提出的。

援助换取选票

这检验第五节，且它在一个单一案例中检验三条回避之路中的两条。

发展援助被换取多边机构中的支持。回路抵抗被分离进诸记法，因为它的两侧凭构成就是不同类的，而第二条回避之路因此在其内失败。而回路在文献中被例行地货币化，援助的价值被定价，而选票的价值被”援助所索取之价格”来估计，因而第一条回避之路被公开地施行，且可被观看着失败：选票的价格从援助被推导，援助的充分性随后相对于选票的价格被评估，而圈子被不加评说地闭合。

被拒绝的礼物与被接受的礼物

这检验第六节与第七节。

两份外交礼物，在每一个物质方面结构上同一：相同的价值、相同的场合、诸方之间相同的关系。第一份作为侮辱被拒绝。第二份被接受。

在一个物质计账上两个案例是相似的，且没有账簿区分它们。差异在于转换是否被接受，也就是说，在于接受方是否同意了”一件一类之物据以算作对一件另一类之物之结算”的条款。

什么会证伪它。倘若接受与拒绝留下无持续的后果，以至于关系无论哪一个发生了都相似地进行，那么结算就是附带现象，而第六节是空转的。

可得退出之下的不利比率

这检验第七节的机制，且它是那一节以之作为赌注的案例。

预测是，一个国家对一项不利结算的接受，会追踪它为承认为一个行动者而对关系的依赖，且会仅弱地追踪它的物质依赖以及物质替代方案的可得性。两个物质上相似、替代方案相似的国家，应在它们会接受什么上依”其作为诸方之地位有多少由所争议的那段关系构成”而分化。

什么会证伪它。倘若接受被证明为被物质依赖完全解释，那么机制就是多余的，且应被丢弃。本文中几乎没有别的东西取决于它。

作为维系性回路的冷战

这检验第四节所断言的不相容，且它被纳入，是因为它是现实主义最满怀信心地主张为己有的案例。

按现实主义的准则该案例是一次成功：一个未坍缩的体系，一种双方持续劳作以保持均等的能力分布，一个跨四十年被维持的均衡。此处所推进的框架预测，在同一时期，每一集团的独立生成能力，在两者被锁定于其中的那一领域以外的每一领域中，相对于一个锁定缺席的反事实，都下降或停滞。

检验是艰难的而反事实是困难的，且这一点不被掩盖。它被纳入，是因为一个对现实主义最满怀信心地主张的案例几乎无话可说的框架，会在论证被交锋的那一点上推却论证。

§12 结论

论证

权力不是一个原初项。它从生成性差异化推出，而推导它而非接受它改变了可以关于它被追问的东西。一种把权力当作给定的理论，大体被局限于追问有多少权力。一种从生成推导出权力的理论，可以追问一次给定的权力是否服务于它由之而来的那个生成。

不对称不是不义，且它抵抗那种在移除它之后仍让生成存续的清除。这是埃格拉什的洞见，而此处所补充者是它所缺的那个本体论之据。对此再加上一个世界政治的诸关系理论未从它们与本文共享的前提中得出的推论：关系本体论并不递交和谐。它递交权力的必然性，并随之递交治理它的必然性。

在一个承载异质价值的回路中，闭合是一个判断而非一个事实。某物必须被迫算作对另一种物的偿付，而它据以被如此迫使的条款并非由价值本身供给。比率是被设定的，且不存在一个正确的。

而在一方若不退出关系便不再算作一方之处，抵抗的强形式对它是不可得的。它无从拒绝被结算。测度化于是不再是一条有争议的边界，而成为关系的一个构成性条件。

国际的情形与其说是测度化社会学所描述过的那个现象的又一实例，不如说是它的极限，在此极限上该现象改变了其性质。

现实主义，最后，是一种结算体制。它把承认、正义、身份与文化转换成单一的标量，并把该转换呈现为对何者驱动国家的描述，而它更宜被理解为一项关于它们的账目据以被保有之条款的规定。

未被确立的诸提议

那个据以宣布诸等价的位置在结构上是那个事物为之而算数的位置。机制解释了替代方案所不解释的东西，而那是一个去认真考虑它的理由而非一个演证。

规范性条件，在没有真实比率的情况下，是比率的可争议性。该提议有一个阈值问题和一个俘获问题，且它在一个本应容易的案例中令人不适。

连络的数学不止是关于该问题之结构的一个有暗示力的类比。尚未被表明该形式化许可任何”以言语进行的论证所不许可”的推论。

被留置开放的诸问题

在退出被封闭时谁来重组回路。生成性正义理论回答说生成者自己做它；在国际体系中他们很少处于能够如此的位置，而本文没有别的答案。它是核心的未解决问题，且它关涉生成性正义是否经受得住它的转置。

一个在一层是生成性的回路是否在低一层隐藏着攫取。框架最终必须是多层次的。它此处不是一个。

多少可争议性够了。未给出任何阈值。

论证的压缩

生成性正义要求价值回到那些生成它的人那里。在价值是异质之处，回报必须跨越诸记法被结算，而结算需要一个不可从它所结算之物推导的比率。

测度化社会学已表明这样一个比率的设定乃是权力的一种样式，且对它的寻常抵抗是彻底拒绝该尺度。

本文关乎那个拒绝不可得的情形：其中一方无从拒绝被结算，因为拒绝就是不再算数。

在那里，正义的条件不是比率是公平的，因为没有公平的比率，而是它仍对被重开开放。

在纲领内的位置

两点评说，不加夸张地提出。

在本文构成其一部分的那个更广纲领之内，该论证是一个别处被工作出来之结构的政治实例。异质价值跨越一段”无一方能从中退出”之关系的结算，形式上接近一段亲密之内异质礼物的结算，而该对应是一个定理而非一个隐喻。它此处未被发展。

在纲领的治理分支之内，本文供给一条设计准则，即比率的可争议性，并陈述那个”若该准则要不止于一个愿望便必须被克服”的障碍。障碍是退出的被封闭，而它尚未被克服。

本文以它所不做之事作结，如它必须。它已论证一个不能离开一段关系的一方很难拒绝它被结算所据以进行的条款，而它剩下的凭借是争议它们。它尚未表明这样一方如何获得争议将据以可能的那个地位、那个语言与那个论坛。那就是那个问题，而它被留置开放，因为它是开放的。

参考文献

Amin, Samir. Unequal Development: An Essay on the Social Formations of Peripheral Capitalism. New York: Monthly Review Press, 1976.

Anderson, Elizabeth. Value in Ethics and Economics. Cambridge, MA: Harvard University Press, 1993.

Bachrach, Peter, and Morton S. Baratz. “Two Faces of Power.” American Political Science Review 56, no. 4 (1962): 947–952.

Bourdieu, Pierre. Language and Symbolic Power. Cambridge, MA: Harvard University Press, 1991.

Broome, André, and Joel Quirk. “Governing the World at a Distance: The Practice of Global Benchmarking.” Review of International Studies 41, no. 5 (2015): 819–841.

Chakrabarty, Dipesh. Provincializing Europe: Postcolonial Thought and Historical Difference. Princeton: Princeton University Press, 2000.

Cox, Robert W. “Social Forces, States and World Orders: Beyond International Relations Theory.” Millennium 10, no. 2 (1981): 126–155.

Dahl, Robert A. “The Concept of Power.” Behavioral Science 2, no. 3 (1957): 201–215.

Deleuze, Gilles. Difference and Repetition. Translated by Paul Patton. New York: Columbia University Press, 1994 [1968].

Eglash, Ron. “An Introduction to Generative Justice.” Teknokultura 13, no. 2 (2016): 369–404.

Emmanuel, Arghiri. Unequal Exchange: A Study of the Imperialism of Trade. New York: Monthly Review Press, 1972.

Espeland, Wendy Nelson. The Struggle for Water: Politics, Rationality, and Identity in the American Southwest. Chicago: University of Chicago Press, 1998.

Espeland, Wendy Nelson, and Mitchell L. Stevens. “Commensuration as a Social Process.” Annual Review of Sociology 24 (1998): 313–343.

Fanon, Frantz. The Wretched of the Earth. Translated by Constance Farrington. New York: Grove Press, 1963 [1961].

Foucault, Michel. Discipline and Punish: The Birth of the Prison. Translated by Alan Sheridan. New York: Pantheon, 1977.

Gilpin, Robert. War and Change in World Politics. Cambridge: Cambridge University Press, 1981.

Hornborg, Alf. “Unequal Exchange Is Not Primarily about Monetary Value.” Nature Communications 16 (2025).

Jackson, Patrick Thaddeus, and Daniel H. Nexon. “Relations Before States: Substance, Process and the Study of World Politics.” European Journal of International Relations 5, no. 3 (1999): 291–332.

Keohane, Robert O., and Joseph S. Nye. Power and Interdependence: World Politics in Transition. Boston: Little, Brown, 1977.

Kindleberger, Charles P. The World in Depression, 1929–1939. Berkeley: University of California Press, 1973.

Krasner, Stephen D. “State Power and the Structure of International Trade.” World Politics 28, no. 3 (1976): 317–347.

Lukes, Steven. Power: A Radical View. London: Macmillan, 1974.

Maturana, Humberto R., and Francisco J. Varela. Autopoiesis and Cognition: The Realization of the Living. Dordrecht: Reidel, 1980.

Mauss, Marcel. The Gift: The Form and Reason for Exchange in Archaic Societies. Translated by W. D. Halls. New York: Norton, 1990 [1925].

Mearsheimer, John J. The Tragedy of Great Power Politics. New York: Norton, 2001.

Mennicken, Andrea, and Wendy Nelson Espeland. “What’s New with Numbers? Sociological Approaches to the Study of Quantification.” Annual Review of Sociology 45 (2019): 223–245.

Morgenthau, Hans J. Politics among Nations: The Struggle for Power and Peace. New York: Knopf, 1948.

Nye, Joseph S. Soft Power: The Means to Success in World Politics. New York: Public Affairs, 2004.

Organski, A. F. K. World Politics. New York: Knopf, 1958.

Ostrom, Elinor. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press, 1990.

Polanyi, Karl. The Great Transformation: The Political and Economic Origins of Our Time. Boston: Beacon Press, 1944.

Porter, Theodore M. Trust in Numbers: The Pursuit of Objectivity in Science and Public Life. Princeton: Princeton University Press, 1995.

Qin, Yaqing. A Relational Theory of World Politics. Cambridge: Cambridge University Press, 2018.

Qin, Yaqing. “The Zhongyong Dialectic: A Bridge into the Relational World.” The Chinese Journal of International Politics 17, no. 2 (2024): 206–221.

Radin, Margaret Jane. Contested Commodities. Cambridge, MA: Harvard University Press, 1996.

Raz, Joseph. The Morality of Freedom. Oxford: Clarendon Press, 1986.

Sen, Amartya. Development as Freedom. Oxford: Oxford University Press, 1999.

Simondon, Gilbert. Individuation in Light of Notions of Form and Information. Translated by Taylor Adkins. Minneapolis: University of Minnesota Press, 2020 [1958].

Spivak, Gayatri Chakravorty. “Can the Subaltern Speak?” In Marxism and the Interpretation of Culture, edited by Cary Nelson and Lawrence Grossberg, 271–313. Urbana: University of Illinois Press, 1988.

Sunstein, Cass R. “Incommensurability and Valuation in Law.” Michigan Law Review 92, no. 4 (1994): 779–861.

Walt, Stephen M. The Origins of Alliances. Ithaca: Cornell University Press, 1987.

Waltz, Kenneth N. Theory of International Politics. Reading, MA: Addison-Wesley, 1979.

Wendt, Alexander. “Anarchy Is What States Make of It: The Social Construction of Power Politics.” International Organization 46, no. 2 (1992): 391–425.

Zelizer, Viviana A. Pricing the Priceless Child: The Changing Social Value of Children. New York: Basic Books, 1985.