How to Read This Treatise

Abstract

This treatise applies the geometric method of Spinoza's Ethics to the subject of influence, manipulation, and propaganda. Its central claim is definitional rather than empirical. The literature of influence credits an instrument, whether a technique, a formula, or a practitioner, with a causal power that belongs to the material: to affects already present in the affected person, to conditions of access and repetition, and to the privation of knowledge about causes. Once the vocabulary is fixed, manipulation is shown to be a relation between an influence and the affected person's understanding of the cause of his own action, which makes it diagnosable, measurable, and reversible, and which dissolves the catalogue of tactics that the trade sells in its place. Twelve Definitions and twelve Axioms are established, drawn from verified propositions of the Ethics and the Tractatus Theologico-Politicus. From them the treatise derives twenty-four Propositions with formal proofs, corollaries, and scholia, covering the mechanism of being moved, the material basis of propaganda, the conditions under which manipulation is unstable, and the conduct that follows. Documentary and empirical claims are separated from deductive ones and marked as such, because no demonstration can establish what happened in 1917 and no experiment can establish that a category error has been made. The treatise is accompanied by a documentary dossier on Edward Bernays and a replication audit of the psychology of influence, and it reports verdicts on both, including the claims that fail.

Keywords: Spinoza, conatus, affect, inadequate ideas, propaganda, public relations, Bernays, manipulation, persuasion, replication crisis.

Prolegomenon

This is the complete text of six Parts, with appendices. The Definitions and Axioms of Part I govern the demonstrations that follow. Documentary claims that establish what happened are proved, if they are proved at all, in Part III and Appendix B. Empirical claims that establish a measurement are proved, if they are proved at all, in Part IV and Appendix A. Geometric Parts may cite those findings as illustrations, and they mark the citation so that it is not mistaken for a step in a proof. The three warrants are not substituted for one another.

The scope of the whole work. It treats influence as a case of causation, and it refuses to treat it as a case of wrongdoing. It covers the mechanism by which one person's power of acting is increased or diminished by another, the material conditions under which that mechanism operates at scale, and the conduct that follows for a person who wishes to be less moved and more nearly the cause of his own motion. It treats the modern influence industry as a body of claims to be audited rather than a body of knowledge to be applied.

What it does not cover. It does not teach techniques of influence. It contains no scripts, no formulas, and no procedures for moving others, and this is a consequence of its argument rather than a concession to taste: if Proposition I is correct, such procedures cannot be transferred as procedures. It does not adjudicate the moral status of particular practitioners, and it does not treat moral vocabulary as a source of causal explanation. It does not claim that human beings are free of influence, because that claim is incoherent, and it does not claim that influence is omnipotent, because that claim is false and is sold to the credulous. Where the evidence is thin, the text says so. Where a widely repeated claim was found to be unsupported by the documents, the text reports the correction and names the source of the error.

The method. Three kinds of claim appear in this work and they are never confounded. Deductive claims are established by demonstration from the Definitions and Axioms, in the manner of the Ethics, and are valid or invalid. Documentary claims are established by archives, dated publications, and primary testimony, and are true or false, and are always open to revision by better evidence. Empirical claims are established by measurement, replication, and meta-analysis, and are reported with their effect sizes and their replication status. The manipulation industry's own mistake is repeated in reverse by many of its critics, who refute a historical claim with a statistical argument or a statistical claim with a philosophical one. Each claim is tried in its own court.

The reader. No prior knowledge of Spinoza is assumed, and no mathematics is used. The reader is assumed to be capable of following a proof, to be willing to hold a conclusion in suspension until it is derived, and to prefer an account of causes to an account of villains. The reader who wants the villains will find them in Part III, where the documents are, and will find that the documents are less flattering to the legend than the legend is to itself.

A note on the text. Citations of the form E4P7 refer to Part IV, Proposition 7 of Spinoza's Ethics; E3D1 refers to a Definition in Part III; Definition of the Emotions XIII refers to Spinoza's numbered list of affects at the end of Part III. Every such reference in this work was checked against the text before it was used. Related propositions may be consulted by a reader who wishes to test the chain.

Chapter 1

The Confusion of Tongues

Every book in this field fails at the first step, and the failure is definitional rather than empirical. This Part fixes the vocabulary. It does not report findings, because no finding is possible until the words are distinguished. The Definitions and Axioms established here govern every Proposition in the five Parts that follow.

"Most writers on the emotions and on human conduct seem to be treating rather of matters outside nature than of natural phenomena following nature's general laws. They appear to conceive man to be situated in nature as a kingdom within a kingdom: for they believe that he disturbs rather than follows nature's order, that he has absolute control over his actions, and that he is determined solely by himself."

Spinoza, Ethics, Part III, Preface

Every book written about influence fails at the first step, and the failure is definitional rather than empirical. The words are used as moral labels and then treated as causal categories, and the substitution is never announced. Nothing that follows from a conflated term can be tested, refuted, or acted upon, which is precisely why the conflation survives. It is not an accident of sloppy writing. It is the product.

Consider what a single act may be called. A physician explains to a frightened man why he must take a drug, and the man takes it. An advertiser repeats a claim about a drug until a bored man buys it. A state broadcasts a reason to enlist until a young man enlists. A friend exploits a confidence to extract money. A teacher installs a habit of reading. A platform orders a feed so that a particular mood is produced in a hundred million people who never learn that the ordering was determined. Six acts, one vocabulary, and the vocabulary assigns them to praise or blame before anyone has asked what happened.

There are six distinct questions hidden in this single word, and each of them can be answered separately:

What was done, in the way of causes. To whom it was done, and in what condition that person already was. What the effect actually was, measured rather than asserted. By what right it was done. What the affected person understood about the cause of his own action. Under what material conditions it was done, including who controlled the channel, the repetition, and the alternative.

The literature answers the fourth question, which is moral, and reports the answer as if it were the second and the third, which are causal. That is the whole trick. A technique is called dark because the practitioner is disliked, not because the technique does anything that an admired practitioner does not also do. The same hundred repetitions are instruction when a school does them and manipulation when a brand does them. The same appeal to fear is public health when the speaker is trusted and propaganda when the speaker is not. Under examination, the labels track the identity of the agent and the interests of the labeller. They do not track mechanism.

Spinoza's method begins with deflation of exactly this kind. The Tractatus Theologico-Politicus opens by distinguishing prophecy, knowledge, and imagination, because the political disputes of his century were conducted with a word that carried three meanings at once. The Ethics opens with eight definitions before a single assertion is made, and the definitions are not preliminaries to the argument. They are the argument's first act. Spinoza distinguishes three kinds of knowledge, and the first of them is opinion or imagination, the knowledge of things from signs, report and chance experience (E2P40S2). To define is to move an idea out of that first kind. A definition that does not distinguish its object from everything else is not a rough tool awaiting refinement. It is a defect, and defects propagate.

The confusion of tongues is not innocent, and it is not merely academic. It is the merchandise. A trade that sells power over other people cannot sell the true mechanism, because the true mechanism is thin: a few affects, repetition, and access to channels. It can sell far more if the mechanism is left vague, because vagueness permits the buyer to project onto the seller whatever power the buyer fears or desires. The seller of influence therefore has a direct interest in the conflation of the six questions. His product is the confusion, and the techniques are the packaging.

Three consequences follow, and they organize the rest of this book. First, since the vocabulary is moral rather than causal, the field has no way to be wrong. A failed attempt is explained as bad execution and not as false theory. Second, since the categories do not distinguish causes, the literature substitutes catalogues of anecdotes for a mechanism, which is why the catalogues proliferate and why none of them converges. Third, the person who wishes to be less moved has been given nothing to work with, because moral labels cannot be converted into causal knowledge. He can only become more suspicious, which is itself an affect, and which is itself sellable.

The remedy is not more suspicion. It is definition. What follows in this Part is an attempt to give the subject the vocabulary Spinoza would have given it: a small set of terms that name causal relations, and no terms that name sins.

Chapter 2

The Free Man Who Is Not

The popular defence literature rests on an impossibility. It promises a condition in which a person is no longer subject to influence, and it teaches a set of techniques by which this condition is supposed to be reached. There is no such condition, and the proof does not require a laboratory.

From a given definite cause an effect necessarily follows (E1A3). Nothing in the universe is contingent, but all things are conditioned to exist and operate in a particular manner by the necessity of the divine nature (E1P29). A human being is a mode, which is to say a modification of substance determined by other modes, and the idea of him is not the idea of a thing that determines itself alone. It follows that no man is the adequate cause of his own existence. What follows for influence is a second step, and it is not the step that everything is influence. A finite mode is affected in its power of acting by other finite modes (E4P2, E4P4), and whatever affects its power of acting falls under Definition I. No man is therefore exempt from influence, not because all causation is influence, which Definition I denies, but because the active power of a finite mode is always partly determined from outside it.

Against this, the defence literature posits a faculty that stands outside the order of causes and vetoes it. The reader is told to be vigilant, to spot the technique, to refuse the frame, to notice the anchor. Every one of these instructions presupposes a power of decision that is not itself caused, because if the decision were caused then the technique would simply have to act on its causes instead, and the instruction would be useless as a guarantee. The presupposition is the one Spinoza identified as the root of the whole error: men believe themselves free because they are conscious of their desires and ignorant of the causes that determined them (Part I Appendix; Part III Preface; Part IV Preface). The defence literature is built on that ignorance and paid for by it.

The empirical form of the same error is easier to see. Suppose a man is told that a claim is false, and given proof. Does the proof disarm the affect? Spinoza is explicit: a true knowledge of good and evil cannot check any emotion by virtue of being true, but only in so far as it is considered as an emotion (E4P14). An emotion can be controlled or destroyed only by another emotion contrary to it, and with more power for controlling emotion (E4P7). To supply a true idea is therefore not the same as to supply a stronger affect, and by E4P14 the true idea checks an emotion only in so far as it is itself an emotion. The reader who has been told this twice and has forgotten it in the presence of the fear has demonstrated the point rather than refuting it.

Two errors follow from denying this, and they are opposites. The first is the promise of immunity, which sells vigilance as though vigilance were a cause rather than an effect. The second is the paranoid inversion, in which a hidden agent is credited with the power of the entire order of causes, so that every mood, preference, and purchase becomes the work of a manipulator. The paranoid reading is not the opposite of the naive reading. It is the same reading with the sign reversed: in both, an agent, whether oneself or a conspirator, is credited with a causal power that belongs to the whole. There is no conspiracy, only causes. The fear of being manipulated is itself a market, with sellers, and it recruits the same affects that the manipulators use.

What is possible is not exemption but proportion. A man acts when something takes place of which he is the adequate cause, and is acted upon when he is only a partial cause of it (E3D2). Freedom, in the only sense Spinoza would allow for a mode, is the increase of that proportion: more of what one does following from one's own adequate ideas, less from ideas whose causes one cannot state. This is a matter of degree, it is reversible, and it increases where the causes of adequate ideas are present. It cannot be completed, and the man who claims to have completed it is either mistaken about himself or selling something.

This Part fixes the terms. Part II derives the mechanism, because a mechanism can be derived and a rumour cannot. Parts III and IV turn to the documentary and empirical record, where the claims of the trade are tested against what happened and what replicates. Part V returns to deduction and rebuilds the account that survives. Part VI states what follows for conduct.

Chapter 3

Definitions

Definition I. Influence. By influence I understand any operation of one mode upon another whereby the active power of the second is increased or diminished, aided or constrained (E3D3).

Note. The definition supplies an operational clause and nothing more. Whether the word names anything beyond causation, and what it adds to an explanation, is not settled here and is the business of Proposition I. What the definition marks is that the effect in question falls on the active power of the affected thing rather than on its position or its quantity. A stone struck by another stone is caused but not influenced. A man whose appetite is produced by the smell of bread is influenced, because his power of acting is altered.

Definition II. Adequate and inadequate cause. As Spinoza: by an adequate cause I understand a cause through which its effect can be clearly and distinctly perceived; by an inadequate or partial cause I understand a cause through which, by itself, its effect cannot be understood (E3D1).

Note. The distinction is the spine of this book. Everything that will be said about manipulation, about freedom, and about the difference between teaching and propaganda is a consequence of it. Note what the definition does not contain: it says nothing about intention, deception, or harm. Two men may both be partial causes of an effect, and one may have intended it. Adequacy is a property of the causal relation, not of the moral quality of the agent.

Definition III. Action and passion. As Spinoza: I say that we act when anything takes place, either within us or without us, of which we are the adequate cause; and that we are acted upon when something takes place in us, or follows from our nature, of which we are not the adequate cause but only a partial cause (E3D2).

Note. The words action and passion are therefore technical terms with degrees, not states. No man is entirely one or the other, and no man is either one all the time. The practical question in every case is quantitative: what proportion of this act can be understood through the agent's own nature alone, and what proportion depends on causes outside him. The measure is the intelligibility of the effect through the agent, not the agent's ability to recite causes, since by Axiom V he is a poor witness to his own.

Definition IV. Manipulation. By manipulation I understand influence in which the affected person's idea of the cause of his own action is inadequate. The relation holds in the same degree as that inadequacy.

Note. This definition does the work that a catalogue of tactics pretends to do, and it does more of it. It is a relation between the influence and the state of knowledge of the person influenced, not a property of the technique. It explains immediately why identical acts receive opposite names depending on who is moved and what he understands: the physician who explains the mechanism produces an adequate idea of the cause, and the advertiser who produces the desire while concealing the mechanism does not. It explains why manipulation scales with the ignorance of the affected rather than with the skill of the agent. It explains why the same technique is trivial against one man and decisive against another, and why a population's resistance is a measure of its causal knowledge rather than of its moral fibre. And it makes manipulation a matter of degree, since ideas of causes are more or less adequate, which is the truth of the matter and which a catalogue of sins can never express.

Definition V. Persuasion. By persuasion I understand influence whose effect is an adequate idea of the cause of the affected action, that is, an idea which is clear and distinct, whether by common notions or by intuition (E2P40S2), the knowledge of an effect depending on and involving the knowledge of its cause (E1A4).

Note. Persuasion and manipulation are distinguished by the kind of idea produced in the affected person, not by the intention of the agent. This is deliberate, and it is the second stroke of the definitional work. Intentions are mostly hidden, frequently from the agent himself, and are never available to the person being moved, so a definition that depends on them is unusable by the only person who needs it. Products, by contrast, are diagnosable, though not by the testimony of the person moved. What a man reports about the cause of his own action is evidence about his imagination and is not a verdict, because the mind does not know itself except in so far as it perceives the ideas of the modifications of the body (E2P23), and those ideas, referred to the mind alone, are confused (E2P28). The test of adequacy is whether the action can be understood through the agent's own nature alone (E3D1, E3D2). His report remains useful in one direction only: a man who can state no cause at all is giving evidence that his idea is inadequate, while a man who can state causes has not thereby proved his idea adequate. It follows that the same agent may manipulate and persuade in the same sentence, and that the question of what he meant is beside the point.

Definition VI. Propaganda. By propaganda I understand influence that is directed by an organized body, on matters of collective action, addressed to a multitude, and sustained by repetition.

Note. Four conditions, all of them causal: organization, collective object, scale, repetition. A lie told once to a friend is a lie, not propaganda. A lie told a million times by an institution with a budget is propaganda, and its power is not the power of the lie. The definition is silent on truth, which is the point. True statements repeated by an organized body on matters of collective action are propaganda in this sense, and a great deal of what the trade calls education, public health, and brand communication falls inside it. The objection that this makes the term uselessly broad is an objection to the facts, not to the definition.

Definition VII. Coercion. By coercion I understand influence that operates by physical constraint, such that the alternative course of action is unavailable rather than merely undesirable.

Note. The distinction between coercion and consent is the practical distinction that every state must manage, and Spinoza's political writings turn on it. A man who hands over his money at knifepoint does not consent, and the state that secures obedience by force does not govern by opinion. Much of what appears as persuasion is the residue of an earlier coercion, or its advertisement. The distinction will be used in Part II to explain when propaganda is produced at all.

Definition VIII. Consent. By consent I understand obedience produced by an affect rather than by constraint.

Note. Consent, so defined, is the state's real business, because force is expensive, visible, and self-limiting. A government that must hold a gun on every citizen is not a government but an occupation. The ordinary case is the one Spinoza describes: men are moved by their affects, and the sovereign who can direct those affects commands without the expense of force. This is why the state takes an interest in opinion, and why the question of propaganda is a question about government rather than about advertising.

Definition IX. Interest. By the interest of an agent I understand the direction of its endeavour as determined by its actual conditions (E3P6, E3P7).

Note. Interest is not a moral category and not a psychological motive in the ordinary sense. It is a statement about what an agent's situation makes it strive for. It is used in this book for one purpose: to explain why an agent's account of his own power is the least reliable evidence about that power. A merchant who tells you that his goods are potent has told you about his situation.

Definition X. The multitude. By the multitude I understand a plurality of men considered in so far as they are determined by affects rather than by reason. Marked extrapolation. The term is Spinoza's and it is political rather than psychological. It is not defined in the Ethics, where the word occurs four times and never as a technical term; its home is the Tractatus Theologico-Politicus and the Political Treatise. The Ethics supplies the premise only, that men rarely live in obedience to reason and are generally envious and troublesome (E4P35S).

Note. The multitude is not a crowd, a mob, or a class. It is the ordinary condition of large assemblies, including those made of educated men, including the readers of books about influence, including the author of this one. Spinoza's political realism consists largely in taking this as the normal case rather than as an aberration.

Definition XI. Custom. By custom I understand a disposition of body and mind such that a later affection of one kind recalls the affect originally associated with it (E3P14), and, where the association has been produced often enough to become characteristic of the body, such that the later affection produces its affect without the original object being present. Marked extrapolation. E3P14 establishes the association of two emotions and E2P18 the association of two bodies, and neither establishes habituation by repetition. The extension of association to custom is Spinoza's own move in the political writings and is not demonstrated in the Ethics.

Note. Custom is not wisdom, virtue, or a moral achievement. It is conditioning, and it is the physical basis of what the trade calls branding, message discipline, and saturation. Because it requires repetition and repetition requires access, custom is expensive to produce, which explains why the entities that own distribution channels win arguments that they would lose on the merits.

Definition XII. Freedom. As Spinoza: that thing is called free which exists solely by the necessity of its own nature and is determined to act by itself alone (E1D7). For a finite mode this is a limit case rather than a condition. What follows for a man is not stipulated here but derived, in Part V.

Note. Freedom so defined is compatible with being caused, because a thing's own nature is itself a cause. What it excludes is being a partial cause of one's own action and mistaking the remainder for one's own doing. The reader who wants a doctrine of exemption should look elsewhere. The reader who wants the only account of freedom that survives contact with causation has it in this definition, and will find it demonstrated in Parts V and VI.

Chapter 4

Axioms

Axiom I. From a given definite cause an effect necessarily follows (E1A3).

Axiom II. The order and connection of ideas is the same as the order and connection of things (E2P7).

Note. Influence therefore has two faces that are one face. There is no realm of discourse separate from the realm of bodies, and no explanation of a persuasion that must be stated in psychological terms and cannot be stated in material terms. This axiom is the reason the book will treat distribution, repetition, and access as causes rather than as contexts.

Axiom III. Nothing in the universe is contingent, but all things are conditioned to exist and operate in a particular manner by the necessity of the divine nature (E1P29).

Axiom IV. Everything, in so far as it is in itself, endeavours to persist in its own being. This endeavour is nothing else but the actual essence of the thing (E3P6, E3P7).

Axiom V. Men are conscious of their desires and ignorant of the causes by which they are determined (Part I Appendix; Part III Preface; Part IV Preface).

Note. Spinoza's observation is precise and its consequences are not flattering to either side. The manipulator is not exempt: he too is conscious of his striving and ignorant of what disposed him toward it. The axiom is used in Part III to explain why a propagandist's account of his own power is biased, and insufficient without the comparison Proposition I requires.

Axiom VI. The ideas of the modifications of the body, considered only in relation to the mind, are not clear and distinct but confused (E2P28); falsity consists in the privation of knowledge which inadequate ideas involve (E2P35); and inadequate and confused ideas follow by the same necessity as adequate or clear and distinct ideas (E2P36).

Note. The last clause is the most important sentence in this book after Definition IV. It forbids the conspiratorial reading of manipulation. A crowd moved by a false account of its cause is not a violation of the order of nature, a lapse, or an evil. It is an instance of nature operating, and it was as necessary as any demonstration. Nothing is more misleading than the assumption that confusion is an anomaly.

Axiom VII. An emotion can only be controlled or destroyed by another emotion contrary thereto, and with more power for controlling emotion (E4P7).

Axiom VIII. A man is as much affected pleasurably or painfully by the image of a thing past or future as by the image of a thing present (E3P18).

Note. This axiom is the entire mechanical basis of advertising, insurance, prophecy, threat, and hope. It also explains why reasoning with a frightened man about a possible future is not equivalent to reasoning with a calm man about the same future, since the image acts on him as though the thing were here.

Axiom IX. We endeavour to do whatsoever we conceive men to regard with pleasure, and shrink from what we conceive men to shrink from (E3P29); and by the very fact that we conceive a thing, which is like ourselves, and which we have not regarded with any emotion, to be affected with any emotion, we are ourselves affected with a like emotion (E3P27).

Axiom X. The knowledge of good and evil is nothing else but the emotions of pleasure or pain, in so far as we are conscious thereof (E4P8).

Note. Moral vocabulary is therefore a report of affect in the guise of a description of the world. When a man says that a practice is evil, he has told you that the practice produces pain in him and has told you nothing about the practice. The argument of Section I of this Part is an application of this axiom.

Axiom XI. An emotion which is a passion ceases to be a passion as soon as we form a clear and distinct idea thereof (E5P3); and there is no modification of the body whereof we cannot form some clear and distinct conception (E5P4).

Axiom XII. There are as many kinds of pleasure, of pain, of desire, and of every emotion compounded of these, such as love, hatred, hope, fear, as there are kinds of objects whereby we are affected (E3P56).

Note. The heterogeneity of affect is an invalidation condition of the entire genre of technique books. Any single script, formula, or pattern must act on different objects in different men, and the objects are not the script. What a technique can do is select for the men whose objects it happens to fit. Selection is the mechanism that the literature reports as persuasion, and Part IV will show the measurements that distinguish them.

Chapter 5

Proposition INo technique of influence has power apart from the conditions under which it is a cause.

By technique, in the Propositions that follow, I understand a named form of words, gesture, or script offered as a portable cause of influence, considered apart from the set of conditions sufficient for the effect.

Depends on: Definition I, Axiom I.

Proof. By Definition I, influence is an operation of one mode upon another whereby the power of acting of the second is changed. By Axiom I, that change, being an effect, follows necessarily from a definite cause. A technique, as just defined, is a named form offered as that cause, considered apart from the sufficient set. Considered apart from the set, it is not the definite cause, and therefore, by Axiom I, the effect does not follow from it. Considered as a member of the set, it has power only as one condition among those that jointly suffice, which is to say it has no power apart from the conditions under which it is a cause. What such a named form can do, when it does anything, is alter the set: it removes a condition whose presence had been preventing the change, or it adds a condition whose absence had been preventing it. If it leaves the set unaltered, then by Axiom I the change either was already following or continues not to follow. The name technique conceals which of those operations, if either, is at work. Nevertheless.

∎

Corollary. Wherever a technique appears to succeed with neither a change of condition nor the removal of a countervailing cause, either the effect was already in progress and will be attributed to the technique, or the report of success is mistaken.

Corollary II. The measurement of a technique's power requires the comparison of the same act under changed material conditions, since the technique's contribution cannot be separated from the conditions by any inspection of the act itself.

Scholium. The reader may object that this is a verbal manoeuvre, since a technique that removes a countervailing cause has done something. It has. The point is not that techniques do nothing. The point is that what they do cannot be understood, predicted, or transferred unless it is described as one of the two operations above, and the trade describes it as neither. Consider what is actually happening when a salesman's script appears to work. The script has not created a desire. The desire was already there. The script has not supplied information, or it has supplied information and the customer was already willing to act on it. The script has operated on one of four things: it has made a desire salient, it has attached a desire to a particular object, it has interrupted a countervailing affect such as the fear of spending, or it has been delivered under conditions of scarcity, authority, or social proof that constitute changes in the customer's situation. Every one of these is either a change of condition or the removal of a countervailing cause, and none of them is the script.

The consequence is practical and it is expensive for the trade. A technique cannot be packaged, because it is not a thing. What can be packaged is a description of conditions. This is why the honest residue of the influence literature is so much smaller than the literature: the working part of it consists of statements about situations, repetition, and access, and those are not saleable as secrets. The dishonest part consists of scripts, and the scripts work in proportion to the conditions that their buyer already controls without knowing it.

The empirical prediction of this Proposition is testable and Part IV will test it. If influence is mediated by conditions rather than carried by techniques, then the measured effects of persuasion should be small, heterogeneous, and highly sensitive to context, and the celebrated findings should fail to replicate once the conditions are held constant. Whether the record matches that shape is an empirical question, marked [E] where it is answered.

Chapter 6

Proposition IINo influence creates the striving. It determines an endeavour already present to a new object.

Depends on: Axiom IV, Axiom XII, Definition of the Emotions I, E3P9, E3P12, E3P15.

Proof. By Axiom IV each thing endeavours to persist in its own being, and this endeavour is its actual essence. By E3P9 the mind, in so far as it has both clear and confused ideas, endeavours to persist, and is conscious of this endeavour; by E3P12 it endeavours to conceive those things which increase or help its power of acting in the body. Desire itself is therefore not an import but the thing itself, in so far as it is conscious of its striving; and every particular desire is that same striving determined to a particular activity by some given modification of itself (Definition of the Emotions I). By Axiom XII the kinds of desire are as many as the kinds of objects, so a determination to a new object is a determination of that same striving, not the creation of a new essence. What an external cause can do is determine that existing endeavour to a particular activity. That such a modification can come from outside is certain, since anything can accidentally be the cause of pleasure, pain, or desire (E3P15), and it follows that a particular desire can arise under an external influence. What cannot arise from outside is the striving itself, which is the essence of the thing, Axiom IV. Nevertheless.

∎

Corollary. The influence that appears most powerful is that which attaches itself to a desire already in force, and the appearance of power belongs to the desire and not to the agent.

Corollary II. An apparent creation of desire is either the redirection of an existing desire toward a new object, or the attribution to an external object of an affect whose cause lies elsewhere in the affected person's situation.

Scholium. The psychology of the last thirty years supplies the demonstration that the old literature could not, and it supplies it by accident. Subliminal cues bias choice only among people who are already thirsty, and only for habitual choices. Priming of the large kind sold in the trade does not hold. The single most robust finding in the applied literature, anchoring, is not a technique for producing desire at all; it is a technique for shifting a number. Where the literature reports large effects from tiny interventions, the replication record contains the explanation: the effects were measured on people who were already disposed toward the outcome and whose disposition was not measured.

This Proposition draws the boundary that the trade cannot draw without destroying itself. A man cannot be made to want what he does not already, in some form, want. He can be made to want it under a description he did not supply, and he can be made to want it now rather than later, and he can be made to want it at a price he did not intend. Those are three substantial powers, and a book that denied them would be as false as the books that inflate them. What is denied is the fourth power, the one that is claimed: the manufacture of wanting itself.

The corollary for conduct is already visible. Since influence selects among desires rather than creating them, the question a man should ask about his own case is not what was done to me but what was already in me that made me available to it. That question is answered by causes, and answering it is the movement from Definition IV toward Definition V.

Chapter 7

Proposition IIIManipulation occurs in proportion to the inadequacy of the affected person's idea of the cause of his own action.

Depends on: Definition IV, Axiom V, Axiom VI, E4P5.

Proof. By Definition IV, manipulation is influence in which the affected person's idea of the cause of his own action is inadequate, and the relation holds in the same degree as that inadequacy. By Axiom VI such ideas are confused and involve a privation of knowledge rather than a positive error planted from outside; they follow by the same necessity as adequate ideas, which is why the relation is ordinary rather than a breach of the order of causes. By Axiom V men are conscious of their desires and ignorant of the causes by which they are determined, so that this privation is the ordinary human condition rather than an exceptional one produced by an adversary. The force of the affect that carries the action is a separate matter, and is defined not by our own power but by the power of the external cause compared with our own (E4P5). Necessity of the idea and force of the affect are therefore not the same quantity as the degree of the relation. The relation varies with the inadequacy named in Definition IV. Nevertheless.

∎

Corollary. The same influence applied to two men whose causal knowledge differs is manipulation in one and persuasion in the other, and the difference is a fact about the men, not about the influence.

Corollary II. Manipulation does not require an agent who conceals anything. It requires only a person who acts from an affect whose cause he cannot state. An agent who conceals nothing may therefore manipulate, and a person may be manipulated by no one, by the arrangement of things, or by an institution that never gives him a thought.

Corollary III. Because causal knowledge can be increased without the influence being altered, manipulation is a reversible relation, and the reversal is a change in the affected person rather than in the world.

Scholium. The first consequence of this Proposition is the dissolution of the trade's central image, which is a scene. In the scene, an operator stands outside the causal order and works upon a passive subject, and the technique is what passes between them. There is no such operator, because no man is outside the causal order (Axiom III), and because by Axiom V the operator is as ignorant of the causes of his own striving as the subject is of his. The producer of a campaign is moved by his client, his client by a market, the market by conditions neither understands. What the trade sells is a portrait of the operator as an unmoved mover, and the portrait is false in the same way that every portrait of an unmoved mover is false.

The second consequence is that manipulation is common to the point of being the default. Consider a man who is angry and knows that his anger is caused by another man's insult. He has an adequate idea of the cause, and his anger, though painful, is not manipulation; it is a passion whose cause he can state. Now suppose he becomes angry at a political faction that has been described to him forty times in one week, and suppose his own condition (hunger, humiliation at work, six hours of a screen) is a real cause of the anger and is unavailable to him. He has an inadequate idea of the cause, and by this Proposition he is manipulated to the precise degree of that inadequacy. No one needed to lie to him. The forty repetitions supplied the object, and the rest was already present. This is the ordinary case, and it is why the book's practical Part concerns causation rather than detection.

The third consequence concerns blame, and it reforms it rather than abolishing it. Spinoza does not say that agents are innocent. He denies only the faculty by which the trade explains their action. A man who arranges forty repetitions for a living is a cause of something, and it is legitimate to name him as a cause. It is not legitimate to credit him with a power that belongs to the arrangement, the repetition, and the affected person's ignorance, because a false account of the cause cannot guide anyone's action. The verdicts in Part IV will be issued on this basis: not whether the practitioners were wicked, but whether the power attributed to them was theirs.

Chapter 8

Proposition IVEvery manipulation is unstable in proportion to the causal knowledge available to the affected person.

Depends on: Proposition III, Axiom XI, Axiom VII, E4P14.

Proof. By Proposition III and Definition IV, the manipulation-relation holds in the same degree as the inadequacy of the affected person's idea of the cause of his own action. As that idea becomes more adequate, the relation holds of less of the influence. That is the instability named here, and it concerns the idea of the cause of the action, not the idea of the emotion considered as a feeling. Separately, by Axiom XI, an emotion which is a passion ceases to be a passion as soon as we form a clear and distinct idea of the emotion itself. A man may have a distinct idea of his fear and no idea why he enlisted; the first dissolves the passion as a passion, the second is what Definition IV requires. The two may accompany one another. They are not the same object, and the second is not carried by the first. What follows is not the mastery of the affect: by Axiom VII an emotion is controlled or destroyed only by a contrary emotion with more power, and by E4P14 a true knowledge of good and evil checks an emotion only in so far as it is itself an emotion. The instability of the manipulation-relation is therefore a disposition of a degreed inadequacy, and not a guarantee that conduct will change. Nevertheless.

∎

Corollary. Every durable manipulation depends on the continued unavailability of an account of its own causes, and its first defence is not secrecy but noise, since noise can prevent a cause from becoming clear without any act of concealment.

Corollary II. Because by E4P14 a true knowledge of good and evil cannot check an emotion by virtue of being true, the instability of the manipulation-relation is a disposition and not a guarantee of changed conduct. An adequate idea of the cause of the action lessens the relation. It does not, by being true, displace the affect. Displacement still requires a contrary affect of greater power (Axiom VII).

Scholium. This Proposition is the hinge of the book's practical claims, and its corollaries are the two things most often got wrong about them. The first corollary explains a feature of modern public communication that the conspiracy literature cannot explain: the sheer volume. If concealment were the mechanism, one silence would suffice, and there would be no need for continuous output. A continuous output is what one finds, and its function is not to hide the cause but to occupy the space in which the cause might become clear. Noise is cheaper than secrecy and more reliable.

The second corollary limits the remedy. It will not do to say that understanding sets a man free, if by that is meant that understanding alone corrects an affect. Spinoza says the opposite in E4P14, and the reader who has resolved against a habit by reasoning and then performed it anyway has confirmed the axiom. The remedy in Part VI is therefore constructed accordingly: not information, but the production of a contrary affect whose object is the true cause, so that the passion is displaced rather than refuted, with the conditions of that production stated rather than assumed. Nothing else is available, because by Axiom VII nothing else works.

Chapter 9

Scholium of the Four Questions

A method follows from the Definitions, and it is the method this book will use against every claim about influence in the Parts that follow.

When someone states that something or someone moves people, ask four questions and require answers to all four.

First, what was the affect already present in the affected person? If the answer is that none was present, the claim is false by Proposition II, and no evidence of a technique can rescue it.

Second, what change of condition was introduced, or what countervailing cause was removed? By Proposition I this is the whole of what a technique can be, and a claim that specifies neither has specified nothing.

Third, who controlled the channel, the repetition, and the alternatives? By Definition VI repetition at scale is a constituent of propaganda, and by Axiom II it has a bodily face as well as a mental one, so the question is a question about causes and not about rhetoric.

Fourth, what does the affected person understand about the cause of his own action, and can his action be understood through his own nature alone? By Definition IV and Proposition III this determines whether the case is manipulation or persuasion. His own account is evidence toward the answer and is never the answer, since by Axiom V he is conscious of his desires and ignorant of their causes.

A claim that survives these four questions is worth an experiment. A claim that cannot answer them is a description of a scene from the trade's literature, and the rest of this treatise exists to show what such claims are worth when they are measured and when the documents are read.

Chapter 10

Fear and Hope Are One Machine

Part I fixed the vocabulary. This Part derives the mechanism that the vocabulary makes visible. Influence never creates the striving. It determines an endeavour already present to a new object, in a body already disposed, through an imagination already stocked, under conditions of access, repetition, and force the influencer usually did not make. What remains is to show how those affects are moved: by which images, under which associations, at what scale, and why demonstration does not interrupt the motion. The Propositions that follow are consequences of Definitions I to XII, Axioms I to XII, and Propositions I to IV. They do not redefine those terms.

Hope and fear are not two instruments. They are one instrument read from two sides of a single doubt. Spinoza defines them as follows.

XII: "Hope is an inconstant pleasure, arising from the idea of something past or future, whereof we to a certain extent doubt the issue."

XIII: "Fear is an inconstant pain arising from the idea of something past or future, whereof we to a certain extent doubt the issue."

The two definitions share their object and their condition. The object is an idea of something past or future. The condition is that the issue is doubted to a certain extent. What differs is only the valence: pleasure in the one case, pain in the other. Remove the doubt and the pair collapses. By the note to E3P18, if the element of doubt is removed, hope becomes confidence (Def. Aff. XIV) and fear becomes despair. Confidence and despair are not levers for further movement in the same way, because the issue is settled. The lever that can be pulled again tomorrow is the inconstant pair, and the inconstancy is the doubt.

There is therefore no hope without fear. He who hopes for an issue thereby fears the failure of that issue, since the same doubt that makes the pleasure inconstant makes a pain available from the contrary image. He who fears an issue thereby hopes for its non-occurrence, since the same doubt that makes the pain inconstant makes a pleasure available from the image of escape. The two affects are not sequential and not optional companions. They are the same modification of the body described from opposite sides of an uncertainty. A propaganda that offers hope without naming a fear has not omitted the fear. It has left the fear to be supplied by the audience, which the audience will do, because the object is future and the issue is not known. A propaganda that offers fear without naming a hope has not omitted the hope. It has left the hope to be supplied as the image of whoever or whatever claims to avert the feared thing.

This is the master lever, and the reason is not rhetorical. The multitude, by Definition X, is a plurality of men considered in so far as they are determined by affects rather than by reason. Men so considered do not live in a demonstrated present. They live in an imagined future whose issue they doubt. By Axiom VIII a man is as much affected by the image of a thing past or future as by the image of a thing present. The feared invasion, the hoped-for pension, the promised market, the threatened shame, the advertised life: each of these, once it is an image, works on the body as if it were here. The organized body that can keep that image inconstant, which is to say doubted, can renew the affect without supplying a new object. That is cheaper than force, cheaper than a new desire, and cheaper than a demonstration.

Spinoza did not address advertising, insurance, modern electoral machinery, or the industrial production of doom. Those applications are ours. The mechanism is not. The mechanism is the pair of definitions together with E3P18.

Consider advertising. The object offered for sale is rarely present in the moment of the appeal. What is present is an image: of future possession, of future relief, of future honour, of a future self whose issue is not yet settled. The pleasure taken in that image is hope. The pain taken in the image of going without is fear. The two are one purchase. The copywriter who believes he has created a desire has done no such thing. By Proposition II the striving was already there, as thirst, status, hunger, loneliness, or the endeavour to persist (Axiom IV). The advertisement has attached that endeavour to a particular object and has made the issue of obtaining it doubtful enough to keep the affect inconstant and therefore renewable. A man who is certain he will obtain the thing has confidence, and confidence does not send him back to the advertisement. A man who is certain he will never obtain it has despair, and despair does not send him to the till. The working customer is the one who doubts. Saturation is the practice of keeping that doubt supplied with images.

Consider insurance. The contract is a present exchange of money for a promise. What is purchased is not a present good. What is purchased is an image of a future loss paired with an image of future indemnification. Both images, by E3P18, affect as present. The premium is extracted from a present pain. The policy is sold as a present pleasure. Hope and fear are the product. The underwriter who speaks of prudence has described an affect in the language of Axiom X: the knowledge of good and evil is nothing else but the emotions of pleasure or pain, in so far as we are conscious thereof. The customer who calls the purchase rational has reported that the image of unaffordable loss pains him more than the image of the premium. He has not inspected a probability. He has been affected. Where the fear is already strong, almost no art is required to complete the sale. Where the fear is absent, no art will suffice. That consequence will be Proposition IX. It is already visible here.

Consider prophecy. Spinoza treated the religious case in the Tractatus Theologico-Politicus, and the opening of that book is a description of this machine: men, unable to govern their circumstances by set rules, fluctuate between hope and fear, and in that fluctuation become credulous. We do not quote the TTP as a demonstration. We note the continuity and mark the rest as application. Secular prophecy is the same pair under another name. The market forecast, the guru's call, the crisis newsletter, the demographic warning, the technological utopia: each supplies an image of a future whose issue is doubted, and each extracts a present affect from that image. The prophet's power is not the power of being right. A prophet who settles the issue has, by Def. Aff. XIV and the note to E3P18, converted hope into confidence or fear into despair, and has thereby destroyed the inconstancy that made him necessary. Durable prophecy preserves the doubt. That is why the date of the end of the world, when it is named, is moved, and why the forecast that can be checked is less useful to the forecaster than the forecast that cannot.

Consider politics. An election is an organized uncertainty about a collective future. The candidate is not, in the imagination of the multitude, a present administrator of present files. The candidate is an image of a future condition of the common life: safety or ruin, dignity or humiliation, continuity or rupture. Campaigns that appear to argue about the present are usually arguing about which image of the future will be allowed to work as present. The party that can attach hope to itself and fear to the other has not persuaded anyone of a demonstration. It has occupied both sides of the one machine. The party that offers only a programme, without an uncertain future attached to it, has offered a text to men who are not being moved by texts. Definition X already said what those men are.

Consider doom. The apocalyptic register, religious or secular, is hope and fear at their highest inconstancy. The feared collapse and the hoped-for remnant, the feared climate and the hoped-for transition, the feared war and the hoped-for deterrence: the objects differ, the machine does not. This Part does not settle which of those futures is likely. Likelihood is a later question, and it is not a question this mechanism waits for. By E3P18 the image works whether the thing imaged will occur or not. By Axiom VI inadequate ideas follow by the same necessity as adequate ones. A well-founded fear and an unfounded fear are not two kinds of causation. They are two epistemic conditions of the same affection. The organized body that industrially produces images of catastrophe is not doing a different thing from the organized body that industrially produces images of deliverance. It is working the other side of the same doubt. The multitude that lives on such images is not being informed. It is being kept in the condition in which hope and fear can be renewed.

The master status of this lever follows from the others rather than from a preference. Imitation, treated below, can spread a present affect; it does not by itself supply a future object. Honour and shame can cheaply govern a present action before an audience; they do not by themselves give a multitude a collective future to act toward. Custom can recall a pairing; the pairing still needs an affect to pair. The uncertain future is the object that matches this machine. Propaganda, by Definition VI, is influence directed by an organized body, on matters of collective action, addressed to a multitude, and sustained by repetition. Collective action, insofar as it is to be produced and not merely recorded, often has its issue in the future. A multitude, insofar as it is determined by affects, then doubts that issue. Hope and fear are the names of that doubt as pleasure and as pain. Where the object is an unsettled future, durable propaganda is an operation on this machine, whatever else it also does. Custom, honour, and liturgy can sustain repetition without that object, as Proposition V will say. They are other operations, not this one.

The trade prefers to describe the operation as a technique: a fear appeal, a hope appeal, a vision, a warning. The names conceal the identity. There are not two techniques. There is one condition, uncertainty about a future that has been made into an image, and two readings of the pain or pleasure that image produces. A book that catalogues fear appeals as one chapter and aspiration as another has cut a single definition in half and sold the halves as a method.

Chapter 11

The Image of the Future Works Like the Present

Axiom VIII states the mechanism that the previous section used and did not yet isolate. A man is as much affected pleasurably or painfully by the image of a thing past or future as by the image of a thing present (E3P18). The proposition is not a metaphor and not a claim about error. So long as a man is affected by the image of anything, he regards that thing as present. The image, regarded in itself, is identical whether it is referred to time past, time future, or time present. The disposition of the body is therefore identical. Pleasure and pain do not wait for the calendar.

This is why advertising works when it works, and why the working is not a trick played on a faculty of judgment. The advertisement presents an image. The image affects. The affection is present. The object may be a car that will not be owned until next year, a holiday that will not be taken, a body that will not be inhabited, a retirement that will not arrive. None of that delay appears in the body as delay. Delay appears only when the image is joined to other images that exclude the thing's present existence, and those excluding images are themselves affections, which may or may not be stronger. By Axiom VII an emotion is controlled or destroyed only by a contrary emotion with more power. The future tense of the copy is not a contrary emotion. It is a grammatical marking. The frightened or hopeful body has already been paid.

This is why insurance works when it works. The flood has not occurred. The illness has not been diagnosed. The lawsuit has not been filed. The image of each is sufficient. A salesman of policies who can make the image vivid has not supplied information about a distribution of outcomes. He has produced a present pain and a present pleasure, and the signature is the motion by which the body attempts to persist (Axiom IV) under that pair. The actuarial table, if it is shown at all, is shown to a man already affected. By E4P14 a true knowledge of good and evil cannot check any emotion by virtue of being true, but only in so far as it is considered as an emotion. The table checks the fear only if the table is itself a stronger affection, which it almost never is, because a column of figures is a weak image beside a house under water.

This is why prophecy works when it works. The prophet's object is, by definition of the role, not present. The vision, the forecast, the secret date, the coming kingdom, the coming crash: each is an image of a future. Each affects as present. The believer who sells his goods, the trader who sells his position, the citizen who consents to an emergency power, is not a man who has made a calculation about tomorrow. He is a man in whom tomorrow is already here as an affection. Spinoza's account of prophecy in the TTP treats this as a fact about imagination, not as a fact about fraud. The application of that account to financial prophecy, geopolitical prophecy, and therapeutic prophecy is ours. The mechanism does not change when the robe is replaced by a chart.

This is why threat works when it works. Punishment deferred is still punishment present, if the image is in the body. The state that displays the scaffold, the employer that mentions the list of names, the platform that shows the banned account, the parent that describes the consequence: each is operating Axiom VIII. The alternative to the display is the actual constraint of Definition VII, which makes the forbidden course unavailable rather than merely painful to imagine. Display is cheaper. It is also less reliable, because an image can be displaced by a stronger image, whereas a locked door cannot. The economics of that difference is Proposition VI.

The same axiom explains a fact that the defence literature cannot explain and therefore denies. Reasoning with a frightened man about a possible future is not equivalent to reasoning with a calm man about the same future. The sentence that names the future may be identical. The affection is not. In the frightened man, by Axiom VIII, the image of the feared thing already affects as a present pain. The reasoning is then an operation against a pain that is here, not an operation upon a hypothesis that is there. By E4P14 the truth of the reasoning does not check that pain by being true. By Axiom VII only a stronger contrary emotion checks it. In the calm man the image is weak or absent. The same sentence meets no present pain, or meets a pain too small to govern. The two conversations are not two performances of one technique. They are two different causal situations, and the technique, by Proposition I, was never the cause.

The defence literature treats this as a failure of the frightened man: he would see the point if he were more rational, more informed, less emotional. That diagnosis inverts the order. The frightened man is not failing to reason about a future. He is being affected by a present. Asking him to treat as merely possible what his body already treats as present is asking an affection to yield to a grammatical mood. It will not. What will yield, when anything yields, is the image, and only to a stronger image or to a clearer idea of the affect itself (Axiom XI), which is not the same as a lecture on likelihoods.

A further consequence, which Part I already prepared and which this axiom completes, concerns time. The trade speaks of urgency, scarcity, deadlines, limited offers, last chances. These are not independent devices. They are methods for joining the image of a future good to the image of a future exclusion, so that the present affection is intensified by doubt. The limited offer increases hope by increasing fear of missing the issue. It is the machine of Definitions XII and XIII with a clock painted on it. The clock is not a cause. The doubt is the cause. A man who does not doubt the issue, who is confident he can obtain the thing later or confident he does not want it, is not moved by the clock. A man who doubts is moved, and the movement is present.

Spinoza did not discuss clocks on offers. He discussed images of things past or future. The clocks are our application. They do not add a mechanism. They illustrate one.

Chapter 12

We Want What Others Want, and Want It More

Axiom IX supplies the second machine, cheaper than the first and more constantly available. We endeavour to do whatsoever we conceive men to regard with pleasure, and shrink from what we conceive men to shrink from (E3P29). By the mere fact that we conceive a thing like ourselves to be affected with an emotion, we are affected with a like emotion (E3P27). The two propositions are one fact about the imitation of affects and the desire for approval. A man need not be shown a future. He need only be shown a man.

Honour is pleasure accompanied by the idea of some action of our own which we believe to be praised by others (Def. Aff. XXX). Shame is pain accompanied by the idea of some action of our own which we believe to be blamed by others (Def. Aff. XXXI). These are not moral achievements and not moral failings. They are affections whose external cause is the imagined judgment of others. The imagined judgment, by Axiom VIII, works as a present judgment. The praise need not have been uttered. The blame need not have been uttered. The idea of the utterance is enough. A culture that keeps such ideas stocked, by display, by story, by metric, by the visible faces of an audience, has a lever that does not require a product, a threat, or a future catastrophe. It requires only that men conceive other men.

Envy is hatred, in so far as it induces a man to be pained by another's good fortune and to rejoice in another's evil fortune (Def. Aff. XXIII). This book does not distinguish envy from hatred. The trade that distinguishes them, that praises a healthy form of the one and condemns the other, is making a moral cut that the definition does not support. Hatred is pain accompanied by the idea of an external cause. Envy is that pain directed at another's good. The political campaign that shows the neighbour's gain, the advertisement that shows the rival's possession, the feed that shows the stranger's holiday: each is an operation on hatred under a name that makes the operation look like aspiration. Aspiration, in so far as it is the desire of something engendered by conceiving that others have the same desire, is the imitation named by Axiom IX. The pain taken in the other's having it already is envy, which is hatred. The two arrive together. A literature that sells the first and denies the second has sold half a definition.

Emulation is the same imitation where the object is conceived as honourable, useful, or pleasant. Spinoza gives it a separate name in the catalogue of the emotions. We do not need the separate name, and we do not cite a definition that is not already in force. What matters is the cause, which is E3P27. He who runs because he sees others run, he who fears because he sees others fear, he who buys because he conceives men like himself to have bought: such a one is affected by a like emotion. The organizer of a campaign who can display the affect has saved the cost of producing it. Display is cheaper than argument and cheaper than payment.

These are cheap levers, and their cheapness is the point. Hope and fear require the construction and maintenance of an uncertain future. Imitation requires a visible peer. Honour and shame require an imagined audience. The modern platform is an industrial form of all three at once, and Spinoza did not address platforms; the application is ours. A feed displays the affects of others (E3P27), displays the praise and blame of others (Def. Aff. XXX, XXXI), and keeps a future uncertain by the next item, which may bring honour or shame and is not yet known (Def. Aff. XII, XIII). The user who reports that he is keeping up with news has given an inadequate idea of the cause of his own action (Definition IV). The cause is the imitation of affects under repetition. The news is the object to which that imitation is attached.

Honour cultures, whether they call themselves traditional, professional, military, academic, or online, are not mysterious. They are populations in which the imagined audience is kept continuously present, so that Def. Aff. XXX and XXXI can be operated at low cost. The duel, the rank, the title, the uniform, the citation count, the follower count, the rating: each is a device for making the idea of praise or blame more constantly available to the body. The man who would be moved by a demonstration, if he were considered under reason, is not being considered under reason. He is being considered as a member of a multitude whose agreement, by E4P35, cannot be expected from reason and must be expected, if it is to be expected at all, from a common affect. Honour is that common affect when the common object is the imagined praise of the group. Shame is its pain. The group that can administer both need not administer many arguments.

Sales uses the same cheapness. The testimonial, the queue, the sold-out mark, the room of nodding faces, the phrase that names how many others have already acted: these are not proofs of the product. They are displays of the affects of others, and by E3P27 they produce a like affect in the viewer who conceives those others as like himself. The viewer who does not conceive them as like himself is not moved, which is why testimonials are matched to the target and why a testimonial from the wrong class produces nothing or produces contempt. Heterogeneity of objects (Axiom XII, E3P56) again: the lever is not the testimonial. The lever is the likeness.

State propaganda uses honour and shame when it can, because they are cheaper than hope and fear constructed from nothing. The poster that asks whether one has enlisted, the broadcast that shows the loyal crowd, the ceremony that makes the dissenter visible as a body to be blamed: these are Def. Aff. XXX and XXXI in public. Enlistment campaigns in particular run the two machines together. The image of invasion or defeat is hope and fear (an uncertain future). The image of the man who did not go, blamed by those he conceives as like himself, is shame. The image of the man who went, praised, is honour. The recruit is not created by a slogan. The recruit is a young man already disposed by conatus, by the imitation of his cohort, and by whatever fear of exclusion or hope of glory his situation has already installed. The poster selects among such men. It does not manufacture them. Proposition II has already said this. The present section names the affects that do the selecting.

The desire for approval is cheap for the operator and expensive for the one operated upon, because it places the cause of his pleasure and pain outside him, in an audience he does not command and whose judgment he cannot make adequate. By Axiom V he is conscious of the desire and ignorant of its causes. By E2P28 the ideas of the modifications of his body, considered only in relation to the mind, are confused. He reports that he did what was right, or what was expected, or what anyone would have done. He has named an affect and mistaken it for a reason. The operator who congratulates himself on having persuaded him has made the same mistake from the other side.

Chapter 13

Repetition Builds the Second Nature

Custom is Definition XI. Its ground is E3P14: "If the mind has once been affected by two emotions at the same time, it will, whenever it is afterwards affected by one of these two, be also affected by the other."

This is the whole of branding, the whole of message discipline, the whole of saturation, and the whole of what the trade, when it is being honest for a sentence, calls conditioning. Two affections, once simultaneous, become a single later trigger. The mark and the feeling, the slogan and the fear, the face and the trust, the anthem and the pride, the colour and the party, the notification sound and the hope of honour: any pair will do, and the pair does not ask whether the conjunction was reasoned. By E3P15 anything can accidentally be the cause of pleasure, pain, or desire. Accident here is not chance in the sense forbidden by Axiom III. It is conjunction without an adequate idea of the cause. The second nature is the first nature recalled by a cue.

Cleverness is a single affection, or a short series. It may be intense. It is not, by E3P14, a pairing, and it is not, by Definition VI, repetition. A clever message that is seen once has the power of one affection in one body at one time. A dull message that is paired with an affect and repeated has the power of a custom. The trade that sells cleverness is selling what is scarce in the writer's labour and abundant in the writer's vanity. The organized body that wins is the one that can repeat. Repetition at the scale of a multitude is not a literary property. It is access to bodies, many times. Access is owned, rented, or seized. Capital is the ordinary name of that ownership. Spinoza did not treat capital. The application of E3P14 to the ownership of channels is ours, and it follows from Axiom I and Axiom II without remainder: the effect of repetition follows from the definite cause that produces the repeated affection, and the order of those ideas is the order of the things that carry them.

This is why message discipline beats improvisation. Discipline is the refusal to introduce a third affection that would fail to recall the pair, or that would pair the mark with a contrary affect and begin a rival custom. The committee that kills the interesting variation is not failing aesthetically. It is protecting a conjunction. The artist who complains that the campaign is stupid has measured the sentence by a standard that is not the cause of the effect. The cause of the effect is the tenth pairing, not the first elegance.

This is why saturation beats argument. An argument, even a true one, is an affection among others, and by E4P14 it does not check a contrary emotion by being true. Saturation is the production of so many pairings that the cue recalls the affect before any argument is encountered, and often before any argument can be formed. The man who finds himself pained by a colour, a font, a name, a jingle, and who cannot state why, is not mysterious. He is E3P14 plus Axiom V: two emotions were once together, one of them has returned, the other has returned with it, and he is conscious of the pain and ignorant of the pairing. The brand is that ignorance given a legal name.

Brand saturation in a city is the visible form of this proposition. The same mark on the board, the vehicle, the bag, the shirt, the screen, the receipt: each impression is an opportunity for the pair to be refreshed. The mark is one of the two affections. The other is whatever pleasure, pain, or desire was present when the mark was first or most strongly paired: thirst, belonging, fear of exclusion, hope of status, the relief of recognition. A man who has never been thirsty, lonely, or ambitious in the presence of the mark will not be a customer of the mark, however often he sees it. Proposition II again. The saturation does not create the thirst. It makes the mark the object to which the thirst, when it occurs, is already joined.

Platform feeds industrialise the same pairing and add a refinement Spinoza could not have named. The feed can pair a mark with an affect and then measure which pairings are followed by further attention, which is a further affection of the body, and then repeat the successful pairs. The refinement is selection, not persuasion. Axiom XII already forbade uniform power across persons: there are as many kinds of pleasure, pain, and desire as there are kinds of objects. The feed that appears to persuade a population has selected, from a heterogeneous multitude, the bodies in which a given pairing already takes, and has then repeated that pairing. Part IV will measure the distinction. The deduction does not wait for the measurement. Selection plus repetition is what a custom is, once the custom can be reallocated by a machine.

Access and capital are the real scarce goods because E3P14 has a material condition. Two emotions cannot be produced at the same time in a million bodies without a channel that reaches a million bodies, and they cannot be produced ten times without a channel that can be used ten times. The writer of a technique book does not own this channel. The platform, the broadcaster, the ministry, the church, the army, the school, the employer, the owner of the wall along the road: these own it, or rent it. What they own is the possibility of custom. What the copywriter owns is a sentence. By Axiom II the sentence is a thing among things, and without the channel it is a thing that does not meet the bodies it would need to pair. The counterfactual is simple and will return as a lemma: the same words, without the channel, are not the same cause, because they do not produce the same affections. To attribute the effect to the words is to violate Axiom I, which requires the definite cause, and the definite cause includes the access.

The school, the liturgy, the drill, the anthem, the pledge, the onboarding, the terms of service accepted every year: these are custom in institutions that do not call themselves advertisers. They work by the same proposition. Two affections, together, often enough, until one recalls the other. The child who stands for the anthem and feels something he cannot name is not being educated in the sense of Definition V. He is being paired. Later, a fragment of the tune will return the affect, and he will report patriotism, or faith, or professionalism, or the feeling that this is simply what is done. He will be conscious of the desire and ignorant of the cause (Axiom V). That ignorance is not a failure of the institution. It is the design, whether or not anyone in the institution can state the design. An institution may pair without a theory of pairing. E3P14 does not require a theorist.

Chapter 14

The Multitude and the Limits of Reason

Definition X: by the multitude I understand a plurality of men considered in so far as they are determined by affects rather than by reason (E4P35S). The definition is not an insult and not a class. It is a description of the ordinary condition of large assemblies, including assemblies of the educated, including the readers of this book, including its author when he is not living under reason, which is most of the time. E4P35S is explicit that it rarely happens that men live in obedience to reason, and that they are generally envious and troublesome to one another, and that they are nevertheless scarcely able to lead a solitary life. The multitude is that fact considered as a political and commercial object.

Demonstration does not move crowds, and the reason is not that crowds are stupid. The reason is that a demonstration produces an adequate idea only in a mind that follows the reasons, and a multitude is not being considered under that description. Even in a single mind, a true knowledge of good and evil cannot check any emotion by virtue of being true, but only in so far as it is considered as an emotion (E4P14). The demonstration that arrives in a crowd arrives, if it arrives at all, as one more affection: a pleasure of being right, a pain of being contradicted, a hope that the speaker will prevail, a fear that he will not, an imitation of the faces nearby. By Axiom VII that affection controls a contrary emotion only if it is stronger. Truth is not a measure of strength. Strength is a measure of strength.

This is why the public information campaign, conceived as the transfer of propositions into a population, fails in the precise way it fails. The propositions may be true. Their truth is not the lever. If the campaign works, it works because the propositions were attached to a stronger affect than the one they opposed, or because they selected the already willing, or because they were repeated until they became custom. The ministry that attributes the effect to the information has made the trade's error in official language. Part IV will measure such campaigns. The geometry already says what they cannot have done: they cannot have checked an emotion by being true.

E4P63 states a further limit, which the moral literature of public life cannot absorb: "He who is led by fear, and does good in order to escape evil, is not led by reason." A great part of what is called civic virtue is this sentence. The man who pays, who enlists, who silences himself, who recites, who votes as his fear directs, in order to escape a feared evil, may produce an action that a reasonable man would also produce. He is not therefore reasonable. He is led by fear. The organized body that congratulates itself on having produced good conduct by warning the population has produced obedience of the kind Definition VIII names consent: obedience produced by an affect rather than by constraint. It has not produced freedom in the sense of Definition XII, and it has not produced persuasion in the sense of Definition V. It has produced a passion that happens to issue in a convenient act. By Proposition IV the manipulation-relation is unstable in proportion to the causal knowledge available. By Proposition X, which this Part will reach, the same obedience is fragile before any stronger cause.

E4P18 adds a quantitative remark that the trade both uses and misunderstands. Desire arising from pleasure is, other conditions being equal, stronger than desire arising from pain. The proof is in the Ethics and we do not repeat it except to note the ground: desire is the essence of the man, pleasure increases that essence's power, pain diminishes it, and the desire born of pleasure is therefore defined by human power together with the power of an external cause, whereas the desire born of pain is defined by human power only. Other conditions are rarely equal. Fear is often cheaper to produce than a vivid pleasure, because the imagination of the multitude is already stocked with threats, and because an uncertain future is already, for a finite mode, a scene of possible loss. A campaign of pure hope, if the hoped-for object is vivid and the doubt is well kept, will outlast a campaign of mere pain, other conditions equal. A campaign of mere pain, if the pleasurable image is weak or absent, will still move men, because pain is an affect and Axiom VII does not require that the winning affect be pleasant. The trade that concludes from E4P18 that one should always sell dreams has ignored the ceteris paribus clause. The trade that concludes that fear is the only reliable instrument has ignored the proposition. Both trades are selling a half.

E4P35 is the political key. In so far only as men live in obedience to reason do they always necessarily agree in nature. The corollary, which Part I already used to define the multitude, is that they do not so live, and therefore do not so agree. A crowd is not a failed seminar. It is a plurality whose agreement, if agreement occurs, is agreement in a common affect, not agreement in a common notion. The organizer who addresses a crowd with a demonstration has mistaken the object. The organizer who addresses a crowd with a common fear, a common hope, a common honour, or a common shame has addressed the object that Definition X names. He may still fail, because the countervailing affects may be stronger (Axiom VII) or absent in the wrong direction (Proposition II). He has at least named the correct kind of cause.

From this it follows that the demand that public speech be reasonable, if by reasonable is meant demonstrative, is a demand that the multitude cease to be a multitude for the duration of the speech. Sometimes a portion of an assembly does so cease, for a time, in so far as some of its members follow reasons. The portion is not the multitude. The multitude is what remains, and what remains is what elections, markets, wars, and platforms actually move. A politics that pretends otherwise is not an ethical politics. It is a false description of its own material.

The limits of reason in a multitude are therefore not a scandal and not a temporary backwardness awaiting education. They are E4P14, E4P7, E4P63, E4P18, and E4P35 read together, under Definition X. Education, in so far as it produces adequate ideas, is persuasion (Definition V) and is real. Education, in so far as it is the pairing of marks with affects in a school that the child did not design, is custom (Definition XI) and is also real. The two must not be given the same name. The first does not scale to a multitude except in so far as the members cease, one by one, to be considered only as determined by affects. The second scales easily, which is why states prefer it, and why the preference is not a conspiracy but an economy of causes.

Chapter 15

Proposition VWhere propaganda has as its object an unsettled future, the lever that matches its duration is an affect of the imagination directed at that future.

Depends on: Axiom VIII, Definition of the Emotions XII, Definition of the Emotions XIII, Definition VI.

Proof. By Definition VI, propaganda is influence directed by an organized body, on matters of collective action, addressed to a multitude, and sustained by repetition. Collective action, in so far as it is to be produced by that influence and not merely recorded after the fact, often has its issue in what is not yet done. Where that is so, the object of the influence is future, or is past in such a way that its bearing on what is to be done remains unsettled, and the issue is doubted to a certain extent. Custom (Definition XI) can sustain repetition of a settled pairing, as liturgy, anthem, and brand mark do; this Proposition does not deny those levers, and does not rank them. It names the lever that matches the unsettled object. By Axiom VIII (= E3P18), a man is as much affected by the image of a thing past or future as by the image of a thing present, so that this unsettled object, once imagined, affects as present. By the Definition of the Emotions XII and XIII, the inconstant pleasure and the inconstant pain arising from an idea of something past or future whose issue is doubted are hope and fear. Those two are one machine, as already shown: the same doubt, read as pleasure or as pain. Where the object is an unsettled future, that machine is the lever that can be renewed without a new object, because the doubt itself is the condition of the inconstancy. Nevertheless.

∎

Corollary. A propaganda that settles its future, by making the issue certain, destroys the inconstancy that made it durable, and must either cease or invent a new uncertainty.

Corollary II. Honour, shame, imitation, and custom may serve a propaganda, and commonly do. They do not replace the lever of this Proposition where the object is an unsettled future; they are other operations, already named, and they can sustain repetition without that object.

Scholium. The wartime enlistment poster, the peacetime insurance advertisement, the party broadcast in the last month before a vote, and the platform prompt that something is happening now and must be answered, are one kind of object. Each places before a body an image of an issue not yet settled: the invasion that may come, the fire that may take the house, the government that may ruin or save, the event that may pass without one's response. Each extracts a present affect. The art, as the trade calls it, is almost entirely the selection of which image will meet which already present fear or hope. The rest is repetition and access.

State propaganda in its ordinary, non-theatrical form is not a lie factory. It is a scheduler of futures. The ministry that publishes a daily threat level, a growth forecast, a public-health scenario, or a warning about a foreign rival is keeping an issue unsettled in the imagination of a multitude. Whether the underlying facts are sound is a documentary question for another Part. The mechanism does not wait on that question. A true warning and a false warning operate the same lever. The difference between them is moral and epistemic. It is not, by this Proposition, a difference of machine.

Advertising of durable brands is the commercial form of the same scheduler. The brand that sells only a present commodity, without an image of a future self, a future membership, or a future loss of face, has to remake the sale at every encounter. The brand that attaches the commodity to an unsettled future (will I still belong, will I still be desirable, will I still be safe) has a reason to be seen again tomorrow. Saturation is then rational, not as decoration but as the maintenance of doubt. Spinoza did not walk through a city of marks. The marks are our illustration, and they illustrate Definition VI as clearly as a gazette of the seventeenth century.

Doom, again, is not a separate species. The organized body that lives by keeping a catastrophe pending is a propagandist in the sense of Definition VI, even when it is right about the risk, even when it is a scientific body, even when it is a church. The definition was silent on truth. This Proposition is silent on truth. What it names is the lever. The reader who is pained by that silence is being affected, which is his right, and he is not thereby refuting the demonstration.

Chapter 16

Proposition VIPropaganda arises where constraint of a multitude is unavailable or self-defeating; it is then the substitute for force.

Depends on: Definition VI, Definition VII, Definition VIII, Axiom VII.

Proof. By Definition VII, coercion is influence that operates by physical constraint, such that the alternative course of action is unavailable rather than merely undesirable. By Definition VIII, consent is obedience produced by an affect rather than by constraint. An organized body that requires a collective action from a multitude (the situation named by Definition VI) must produce that action either by making the alternatives unavailable, which is coercion, or by producing an affect that issues in obedience, which is consent. Consent so produced may be hope, fear, honour, custom, or persuasion; it is not identified with the lever of Proposition V. Coercion of a multitude requires the physical constraint of many bodies at once, or the credible continuation of such constraint. That is a large expenditure of force, visible, and limited by the number and fatigue of those who must apply it. Where coercion is available and does not itself generate a contrary affect strong enough to destroy the obedience sought, the organized body has no need of propaganda, because the alternative is already closed. Where coercion is unavailable, or where the use of force would, by Axiom VII, produce a contrary affect stronger than the obedience sought and would therefore defeat itself, the organized body that still requires the action must produce consent. The durable, organized, repeated form of that production, addressed to a multitude on a matter of collective action, is propaganda. In that precise sense propaganda is the substitute for force. Whether it is cheaper in a given case is a question of expenditure, marked [E] where it is answered, and is not a step in this demonstration. Nevertheless.

∎

Corollary. The appearance of persuasion in public life is often the advertisement of a coercion that remains available, or the residue of a coercion already applied. The alternative is then undesirable because it has been, or could again be, unavailable.

Corollary II. Force that produces a contrary affect stronger than the obedience sought defeats itself, by Axiom VII. That is the sense in which constraint of a multitude can be self-defeating. It is not a moral status.

Scholium. Spinoza's Tractatus Theologico-Politicus treats the state's power as consisting largely in men's affects, and holds that the firmest dominion belongs to the sovereign who has most influence over the minds of his subjects rather than to the one who is merely feared as a tyrant is feared by his own people. We paraphrase that claim and do not offer it as a quotation to be checked against a demonstration. The extrapolation to propaganda as a substitute for force is ours, and it is the content of this Proposition rather than a hidden theorem of the TTP.

The occupying army that must stand a soldier at every door has not achieved government. It has achieved a continuous coercion, and it will look for a cheaper method as soon as it can: a local administration, a school, a radio, a rite, a ration book that can be withheld, a rumour of what happens to those who do not comply. Each of these is a step from Definition VII toward Definition VIII. The last steps are propaganda in the sense of Definition VI. The first steps are still force, or the memory of force. Most historical cases are mixed, and the mixture does not blur the distinction. It shows the substitution in progress.

The liberal state that cannot, without destroying the affects that legitimate it, cage a population for an opinion, still requires collective actions: tax, census, enlistment or recruitment, compliance with health and safety rules, acceptance of a currency, acceptance of an election. Where the alternative cannot be made unavailable, it must be made affectively costly. The public campaign is that making. It is not a failure of the state's nerve. It is the arithmetic of Proposition VI. A state that could cheaply jail every recusant would not spend on broadcasts. A state that spends on broadcasts has told you, without intending to, that jail is costly or self-defeating.

Commercial propaganda, which is advertising at the scale of Definition VI, arises for the same reason. The firm cannot, in ordinary conditions, make the refusal to buy unavailable. It can only make refusal painful or acceptance pleasant in the imagination. Where a firm does make refusal unavailable (a monopoly of a necessity, a tied contract, a workplace that pays in scrip), it advertises less, or advertises only to protect the arrangement from the contrary affects of those who hate the constraint. The advertising that fills a competitive market is the cheap substitute for a power the firm does not have.

Platforms occupy a middle position that Spinoza did not describe. They can make some alternatives unavailable (the banned account, the removed post, the closed payment), which is a fragment of coercion, and they can saturate the rest with images, which is propaganda. The mix is not a new mechanism. It is Definition VII and Definition VI in one owner. The owner who prefers the second to the first is not kinder. He is cheaper.

Chapter 17

Proposition VIIAt the scale of propaganda, access is a necessary part of the definite cause; the art has no power in excess of that conjunction.

Depends on: Definition VI, Axiom I, Axiom II, Proposition I.

Proof. By Definition VI, propaganda is addressed to a multitude and sustained by repetition. An affect produced in many bodies, many times, is a material operation: the same kind of image, carried by things, meeting bodies. By Axiom II the order and connection of ideas is the same as the order and connection of things, so that the mental face of this operation is not a separate realm in which an art could work without the things. By Axiom I the effect follows from a definite cause, which is the whole sufficient set. Words without access do not meet the bodies; a blank channel without a modification of those bodies produces no affection. Access is therefore a necessary conjunct of the definite cause at this scale, and so is some modification capable of meeting an affect already present. Neither conjunct is the cause by itself. By Proposition I what the literature calls a technique is a change of condition or the removal of a countervailing cause, not a power residing in the form. The art practised on the form has no power in excess of the conjunction. The symmetry is exact: take the distribution away and the celebrated practitioner is a man with a sentence; take the modification away and the channel is empty. Nevertheless.

∎

Corollary. Two propagandists of unequal art and equal distribution will not differ in power by the measure of their art, except in so far as the art is itself a change in the conditions of access or a removal of a countervailing affect already present.

Corollary II. Take the distribution away and the celebrated practitioner is a man with a sentence. The sentence may still affect a few, as any speech may. It is not propaganda, by Definition VI, and it is not the power that was attributed to him. Take the modification away and the channel is a blank, which is equally not that power.

Scholium. The history that Part III will read as documents is already intelligible from this Proposition. The men who are remembered as having moved masses commanded, or served, a ministry, a committee, a newspaper chain, a studio, a network, a platform. Their memoirs attribute the movement to their insight. Definition IX and Axiom V together predict that attribution: they are conscious of their striving and ignorant of the causes, and their interest is to be necessary. The insight, if it existed, was a selection among images already likely to meet affects already present. The cause that scaled the selection was the channel.

A salesman in a room, a preacher in a pulpit, a captain before a company, may be skilful in the ordinary sense: he may notice which affect is already in the bodies before him and attach his object to it. That skill is real and small. It is Proposition II applied by a man who can see a face. It does not become a different kind of power when the same man is given a transmitter. It becomes the same skill plus a distribution. The literature collapses the plus into the skill, because the skill can be sold as a book and the transmitter cannot.

Platforms make the collapse easier to see and harder to admit. The feed is the distribution. The creator who attributes his reach to his craft has mistaken the allocation of the feed for a property of his sentences. Remove him from the allocation and the sentences remain, and the multitude does not. The platform that attributes the allocation to the quality of the sentences has an interest in that story, because the story recruits more sentences, which are the cheap input, while the allocation remains the scarce good. Part IV will measure how little the sentences do once the allocation is held constant. This Proposition is the reason to run that measurement.

The military recruiters who saturate a town, the insurance agents who buy the same hour on the same stations for years, the party that owns the only press, the brand that owns the wall along every road to the factory: these are not illustrations of genius. They are illustrations of command over repetition. A rival genius without the wall, the press, the hour, or the town will lose, and will be told that he lacked craft. He lacked a channel.

Chapter 18

Proposition VIIIThe propagandist's account of his own efficacy is determined by his interest and is therefore biased; it is insufficient without the comparison that Proposition I, Corollary II, requires.

Depends on: Proposition VII, Definition IX, Axiom V.

Proof. By Proposition VII access is a necessary part of the definite cause at propaganda scale. By Definition IX the propagandist's interest is the direction of his endeavour as determined by his actual conditions. Those conditions, if he lives by selling his own necessity, dispose him to place the power in himself, or in the art he can sell, rather than in the distribution he commands or rents. A state propagandist on a monopoly channel need not sell replaceability-insurance; the bias is not universal in form, but it is the ordinary case of a man whose livelihood is the credit. By Axiom V men are conscious of their desires and ignorant of the causes by which they are determined. The propagandist is not exempt. He is conscious of the desire to be credited with the effect, and he is ignorant, as other men are, of the causes that disposed him and of the causes that produced the effect. His testimony is therefore the report of an interest under an ignorance. Such testimony is biased, and it is insufficient as evidence of efficacy until the comparison required by Proposition I, Corollary II, has been made: the same art under poorer distribution, the same distribution under poorer art. Nevertheless.

∎

Corollary. The founding texts of the influence trade, being the self-descriptions of men who sold influence, are to be read as evidence of interest, not as evidence of mechanism. Part III will so read them.

Corollary II. The same weakness attaches to the critic who lives by the fear of propaganda. His interest is the potency of the enemy. His account of that potency is not stronger, as evidence, than the propagandist's account of his own.

Scholium. Bernays, whose documents belong to Part III, is the type and not yet the exhibit. What this Proposition requires, before any archive is opened, is a rule of reading. When a man whose income or office depends on being thought potent describes himself as potent, the description has the epistemic status of an advertisement. It may coincide with the truth, as an advertisement may coincide with the properties of a soap. Coincidence is not warrant. The warrant has to come from a comparison the speaker has no interest in supplying: the same art under a poorer distribution, the same distribution under a poorer art, the same words without the channel. Proposition I, Corollary II, already required that comparison. This Proposition says who will not provide it.

The sales manager who attributes the quarter to the new script, the consultant who attributes the election to the message, the growthist who attributes the curve to the copy, the influencer who attributes the conversion to the hook: each is in the situation Definition IX names. The script, the message, the copy, and the hook are the merchandise. The territory, the list, the budget, the allocation, and the already present affects of the buyers are the conditions. Axiom V forbids us to expect that the speaker will know this. It does not forbid us to know it about him.

A caution, which Axiom V also requires: the author of a geometric treatise on propaganda is not outside the same structure. He has an interest in the potency of his own distinctions. His account of that potency is weak evidence. What is not weak is the dependency of the Propositions on the Definitions and Axioms. The reader who can break a step should break it. The reader who believes the author should be believed because he sounds severe has been moved by an affect of honour (Def. Aff. XXX) and has not followed a proof.

Chapter 19

Proposition IXWhere the countervailing affects are strong, no technique suffices; where they are absent, no further operation on affect is required.

Depends on: Proposition I, Proposition II, Axiom I, Axiom VII.

Proof. By Axiom VII an emotion can only be controlled or destroyed by another emotion contrary thereto, and with more power for controlling emotion. By Proposition II no influence creates the striving; a particular desire may be determined from outside, but only as a determination of an endeavour already present. By Proposition I what is called a technique is a change of condition or the removal of a countervailing cause, not a power in a form of words. Suppose, first, that a contrary affect is already present and strong: a hatred of the advertised object, a love of the rival, a shame that forbids the act, a fear that forbids the purchase, a devotion that forbids the enlistment. The incoming operation, whatever its name, controls that affect only if it is itself a contrary emotion of greater power. A technique, being no desire and no affect except in so far as it changes conditions or removes a cause, does not by being a technique become the stronger emotion. Where it does not become that, it does not suffice. Suppose, second, that the contrary affects are absent, and that the desire toward the object is already present. Then no further operation on affect is required to remove a contrary that is not there. Other conditions of the determinate act (access, opportunity, remaining members of the sufficient set) may still be wanting; their absence is not supplied by a script. By Axiom I the act follows when the definite cause is complete. The technique, as a portable form, is not that cause. Nevertheless.

∎

Corollary. The apparent power of a technique is greatest in the middle case, where a weak contrary affect can be removed or a weak supporting affect can be made salient. The trade measures itself on this middle case and reports the measurement as a general power.

Corollary II. Failure against a strong contrary affect is not bad execution. It is this Proposition. The repair is not a better script. The repair, if there is one, is a change in the conditions that produced the contrary affect, or the introduction of a genuinely stronger affect, which is not a script.

Scholium. The insurance salesman in a town that has just flooded needs almost no sentences. The fear is present, the contrary affect (the pleasure of keeping the premium) is weak, and the pairing of the mark with relief is easy. The same salesman in a town that has never seen water, among men whose conatus is presently occupied by other pains and pleasures, will discover that his craft is nothing. He will be told, by the firm, to improve his close. The firm has an interest in that instruction (Proposition VIII). The town has a stronger countervailing life.

Enlistment repeats the demonstration. In a week when an attack is believed to be imminent, the queue forms without a copywriter. In a year when the war is far, the honour of the regiment is weak, and the fear of dying is strong, no poster suffices. The state that needs the men then either pays, which is a change of condition, or conscripts, which is Definition VII, or waits for a shock that will install the fear of not going. The poster is what the state does while it waits, and while it pretends that waiting is a method.

Brand saturation fails in the same way when it meets a contrary love. The man who hates the company because it injured him, or who loves a rival because his cohort does (Axiom IX), is not a case for a better logo. He is a case of Axiom VII. Campaigns that appear to convert such a man have usually waited until the contrary affect decayed for other reasons, or have selected, without admitting it, the man in whom the contrary affect was already weak. Selection, again, reported as persuasion.

Platform feeds, when they cannot move a user toward a given object, do not invent a new desire. They reallocate: they find another object to which an affect already present will join, or they find another user. The appearance of omnipotence is an appearance of inventory. The inventory is large because the population is large and the affects are heterogeneous (Axiom XII). Heterogeneity plus selection mimics creation. Proposition II forbids the mimic to be taken for the thing.

The middle case, where technique appears to shine, is the case the books are written from. A customer already almost willing, a voter already almost aligned, a user already thirsty: a small change of salience, a small removal of embarrassment, a small display of others like himself, and the act follows. Subliminal cues, as Part I already noted, bias already-thirsty people. Priming of the large kind sold in the trade does not hold. Anchoring shifts a number rather than creating a desire. These are not new claims and not measurements invented here. They are the empirical shadow of this Proposition, and Part IV will treat them as measurements. The geometry says why the shadow has that shape: the technique was never the cause, and it is visible only where the causes were already nearly sufficient.

Chapter 20

Proposition XSuccessful manipulation produces obedience without adequate ideas, and therefore produces fragility: the same affect can be re-captured by any stronger cause.

Depends on: Proposition III, Proposition IV, Axiom VII, Definition IV.

Proof. By Definition IV, manipulation is influence in which the affected person's idea of the cause of his own action is inadequate. By Proposition III, manipulation occurs in proportion to that inadequacy. Successful manipulation is therefore the production of an action (obedience, in the political and commercial cases) together with the continuation of that inadequacy. The person acts from an inadequate idea of the cause. By Proposition IV, every manipulation is unstable in proportion to the causal knowledge available to the affected person. In the successful case that knowledge is unavailable, so the instability is not an instability toward freedom. The manipulation-relation remains. Whether the affect remains a passion is a separate question, answered only by a clear idea of the emotion itself (Axiom XI), which is not supplied by the mere absence of an idea of the cause of the act. By Axiom VII a passion is controlled or destroyed only by a contrary emotion of greater power. Any stronger cause that supplies such an emotion, or that attaches the same emotion to a new object, can take the obedience. The success has produced a body that can be moved again by whoever or whatever next meets it with more force. That is fragility, not control. Nevertheless.

∎

Corollary. The organized body that congratulates itself on having secured a population without teaching that population the causes of its obedience has secured a population that a rival, a shock, or a famine can take by the same handle.

Corollary II. Durability of manipulated obedience is durability of the conditions that keep causal knowledge unavailable (noise, repetition, isolation from contrary affects that would include an idea of the cause). It is not durability of a technique.

Scholium. The fragility is visible whenever a crowd changes its object without changing its appetite. The same hope is attached to a new party, the same fear to a new enemy, the same honour to a new mark, the same hatred (envy) to a new neighbour. The operator who held the old object reports betrayal. He was never holding a conviction. He was holding an affect whose idea of its cause was inadequate, and he has been outbid. Axiom VII does not respect incumbency.

State propaganda that produces loyalty without an adequate idea of the state produces a loyalty that a stronger fear, a stronger hope, or a more vivid honour can redirect. This is not a counsel to tyrants to educate. It is a description of why tyrannies that live by noise are surprised by their own crowds. The surprise is Proposition X. The noise that, by Proposition IV's first corollary, occupies the space in which a cause might become clear, also prevents the passion from becoming an action of which the people are the adequate cause. What cannot become adequate also cannot become stable in the only sense of stability this book allows, which is the increase of adequacy (Definition XII). What remains is a passion ready for a stronger passion.

Commercial brands discover the same fragility when a cheaper rival, a scandal, or a fashion takes the honour they had paired with their mark. The customer who cannot state why he bought will not be able to state why he stopped. He will be conscious of a new desire and ignorant of its cause. The brand manager will commission a better story. The story will not be the cause. The stronger affect will be the cause, and the better story will be, at most, the new pairing.

Platforms are engines of this fragility and live by it. A feed that can re-capture attention by any stronger image has no interest in the user forming an adequate idea of why he is here. An adequate idea of the cause of the action would, by Proposition IV, make the capture less stable as a manipulation-relation. A clear idea of the emotion itself would, by Axiom XI, begin to dissolve the passion as a passion. Those are two ideas. The platform has an interest in neither. The business is the production of obedience without adequate ideas: the scroll, the watch, the return. The fragility is the same fact seen from the other side. A rival feed, a stronger shame, a death in the family, a war, a love, can take the same body tomorrow. The platform's answer is more repetition, which is an attempt to keep its own affect the strongest. That is not technique. That is Axiom VII practised with a budget.

The practical corollary for the person who wishes to be less available is not vigilance in the sense the defence literature sells. Vigilance is itself an affect and can be produced by anyone who profits from fear (Part VI will treat this). The practical corollary is the one Proposition IV already named: a clearer idea of the cause, which is necessary and not sufficient, and a contrary affect whose object is that cause, which is what Axiom VII requires. This Part has now named the causes that have to become clear: the uncertain future that is working as present, the imitated affect, the imagined audience, the pairing that has become custom, the distribution that carried them, and the interest of the man who claims he did it with art.

Chapter 21

Lemmas on the Economics of Attention

The Propositions name the mechanism. The following lemmas name the scarcity under which the mechanism is practised. Spinoza did not treat attention as a quantity, nor capital as a condition of repetition. The lemmas are our application of Axiom I, Axiom II, Definition VI, Definition XI, and Propositions I, II, and VII. They are not new affects and not new techniques.

Lemma I. Attention is a scarce affection of the body. A body can be modified in only so many ways at once, and the mind's ideas of those modifications, being the same in order as the things (Axiom II), cannot be indefinitely multiplied in a single duration. To attend is to be affected by one image rather than another. The organized body that commands attention has not obtained a mental substance. It has obtained a present modification, and it has excluded rival modifications for that interval. Because the modification is scarce, it can be sold, rationed, and fought over, which is the ordinary business of publishers, platforms, and ministries. Because it is an affection, it is subject to Axiom VII: it is taken away only by a stronger affection, not by a request that the person pay attention, and not by information about the importance of the object. The phrase "pay attention" is already the economics. What is paid is a present capacity of the body. What is received is an image. The image may be hope, fear, honour, shame, or the mere pairing that will later recall one of these. It is never a neutral opening onto the world. A man who reports that he was merely looking has given an inadequate idea of a passion.

Lemma II. Repetition at the scale of a multitude requires capital, because it requires access, and access is a thing among things. E3P14 names what repetition does once the affections have met the body. It does not supply the meeting. The meeting is distribution (Proposition VII). Distribution of images into many bodies, many times, is rent, ownership, or seizure of channels: paper, towers, spectrum, streets, schools, liturgies, notification privileges, the default position on a screen. These have prices, guards, and laws. A technique book that omits the price has omitted the cause. Message discipline is cheap in the writer's labour and expensive in the channel. Saturation is not a metaphor. It is a budget that has been converted, by Axiom I, into a definite series of affections. The entity that can afford the series can produce a custom (Definition XI). The entity that cannot afford it can produce a remark. The remark may be true. Truth, by E4P14, is not the lever.

Lemma III. What the trade calls persuasion is commonly selection. By Proposition II the striving is already present. By Axiom XII the kinds of desire are as many as the kinds of objects. A message sent into a multitude therefore meets unlike bodies. In some the message joins an affect already in force and an absent or weak contrary. In others it meets a strong contrary and dies (Proposition IX). The sender who counts only the first group reports a rate of persuasion. He has counted a rate of fit. The feed, the mailing list, the lookalike audience, the targeting brief, are machines for increasing the rate of fit by excluding the bodies in which the fit will not occur. That is a change of condition, which Proposition I already allowed as the whole of what a technique can be. It is not the creation of a desire, and it is not the checking of an emotion by a truth. Part IV will measure how much of the celebrated effect remains when the selection is subtracted. The lemma states the subtraction as a requirement of honesty: a claim of persuasion that does not report the prior disposition of the sample has not reported a cause.

Lemma IV. The counterfactual of the same words without the channel is the test that Proposition I, Corollary II, required, and that Proposition VII made specific to propaganda. Hold the sentences constant and remove the access: the wall, the hour, the allocation, the pulpit, the default. If the effect collapses, the cause was not the sentences. Hold the access constant and vary the sentences within the range that still pairs with the same already present affects: if the effect is stable, the cause was not the craft. The trade refuses both tests, because both tests destroy the merchandise. The honest remainder, after both tests, is small and structural: which affect was already there, which contrary was weak, who owned the repetition, and how inadequate the affected person's idea of the cause remained. Those four are the four questions of Part I, now supplied with the machines that make them answerable. Fear and hope as one inconstant pair. The image of the future working as the present. The imitation of affects and the cheapness of honour and shame. Custom as paired repetition. The multitude as the limit of demonstration. The rest is interest, and interest is the weakest evidence of itself.

What follows, in Part III, is not another Proposition. It is a reading of documents, because no demonstration can settle what a man published in 1928 or what a committee did in 1917 to 1919. The documents will be read under the rule of Proposition VIII. What follows in Part IV is measurement, because no demonstration can settle an elasticity. The measurements will be read under the rule of Propositions I, II, VII, and IX. The present Part has done what a Part of this kind can do. It has named the causes. It has not made anyone free. Freedom, by Definition XII, is not the gift of a chapter. It is an increase in the adequacy of one's own action, and that increase, if it occurs, will occur by the only route Axiom VII and Axiom XI permit: a stronger contrary affect whose object is the true cause, and a clearer idea of the passion that had been mistaken for a reason.

Chapter 22

The Creel Committee, 1917 to 1919

This Part is documentary. It offers annotations, not propositions. No demonstration can settle what happened in 1917. Every load-bearing historical claim is marked [H] at its first use in a section. The court here is the archive, the dated publication, and the primary testimony, all of them revisable by better evidence.

The founding document of the industry is a sales pitch by a man whose business was selling his own power. The power, where it existed, was distribution, access to publishers, state training, and capital. The technique was merchandise. That is the claim this Part examines, and it examines it by reading the documents rather than by deriving 1917 from a definition. What Spinoza can supply is a reason to distrust a seller's account of his own potency. What Spinoza cannot supply is a date, a client list, or a circulation figure. Those belong to the record, and the record is thinner than the legend.

Edward Louis Bernays was born in Vienna on 22 November 1891 and died in Cambridge, Massachusetts, on 9 March 1995, aged 103. [H] His mother, Anna Freud Bernays, was a sister of Sigmund Freud. His father, Ely Bernays, a grain merchant, was a brother of Martha Bernays, Freud's wife. Larry Tye states the relation exactly: his mother was Freud's sister, and his father's sister was Freud's wife. Tye calls him a double nephew. The family emigrated in 1892, when Bernays was about a year old. He attended New York City schools and then Cornell University's College of Agriculture, taking a degree in agriculture in February 1912 to satisfy his father's wish that he join the grain trade. Tye quotes Bernays's own verdict: his three and a half years at Cornell gave him little stimulation and less learning. From 1913 to 1917 he worked as a theatrical publicist, promoting clients including Enrico Caruso and the Diaghilev ballet. These facts are load-bearing only as orientation. They do not establish a theory of influence, and the Freud kinship does not establish a causal chain from psychoanalysis to consumerism. That chain is treated later, where the documents that invented it can be named.

What follows is commentary on a career that was, in the first instance, a career of self-description. The books named the trade. The campaigns furnished the anecdotes. The memoir arranged both into a single life whose central figure was always the same man. Historians who later opened the boxes found a less central figure and a more ordinary set of causes: a wartime committee, a partnership, a roster of corporations, a fee high enough to be taken seriously by the men who paid it, and a talent for remaining the source from which later writers copied. The industry inherited the claim and dropped the limit. The task of this Part is to put the limit back where the documents left it.

Annotation. The Committee on Public Information was created as an independent agency by Executive Order 2594 on 13 April 1917. [H] George Creel was chairman. The Secretaries of State, War, and Navy sat as ex officio members. Its functions were to release government news during the First World War, to sustain morale, and to administer voluntary press censorship. Domestic activities ceased after the Armistice of 11 November 1918. Foreign operations ended on 30 June 1919. The committee was abolished by Executive Order 3154 on 21 August 1919. The administrative history is in the National Archives, Record Group 63. These dates are not in dispute. They fix the duration of the largest official propaganda apparatus the United States had yet built.

Annotation. Nicholas Cull, writing for the Public Broadcasting Service, records the scale. [H] The committee deployed advertisers, artists, and newspapermen, an international news service called Compub, and seventy-five thousand Four Minute Men lecturers. The name records the length of the speech, not a cinema monopoly: they spoke in halls, churches, lodges, and elsewhere, as well as in places where a reel was being changed. The apparatus was official, financed, and national. It had a chairman, a record group, and an end date. It is the first fact a documentary account of American propaganda in this century must register, because everything later sold as a private science of the crowd was practised first as a public office of the state.

Annotation. Bernays's place inside this apparatus is less secure than the legend requires. Cull states that Edward Bernays, the future father of public relations, chaired the CPI Export Service. [H] That title rests on one source. The National Archives administrative history of Record Group 63 names the Associate Chairmen as Carl Byoir and Harvey J. O'Higgins. It does not name Bernays. The qualification is not a courtesy. A single secondary attribution, even a careful one, is not the same object as an appointment recorded in the files of the agency. This Part therefore treats the Export Service chairmanship as a claim with a named source, not as a fact established by the committee's own surviving administration. Byoir's later career is better attested in the same institutional sense: he, like Bernays, built a major commercial practice after the war. The committee trained a cohort. It did not, in its own records, crown a father.

Annotation. In 1919, after service with the American Peace Commission in Paris, Bernays returned to New York to apply the committee's methods to business. [H] The Encyclopedia of World Biography and later popular histories repeat this trajectory, and there is no reason to doubt the sequence of cities. The inference that is usually attached to the sequence is another matter. The inference is that the CPI invented the modern public relations industry and that Bernays carried the invention home in his luggage. The first half of that inference is only partly supportable. The committee trained men who then sold persuasion to corporations. It did not originate the trade. Ivy Lee had run corporate publicity in the United States from 1906. [H] Ron Chernow, reviewing Tye in the New York Times Book Review in 1998, is explicit: Ivy Lee and other practitioners of public relations before the First World War can lay greater claim than Bernays to having fathered the craft. The CPI is better described as a catalyst and a training ground than as the sole ancestor of the industry.

The distinction matters because the industry's own origin story prefers a single inventor and a single war. A single inventor can be quoted. A single war can be made to look like a laboratory in which a technique was discovered and then transferred, intact, to soap and cigarettes and bananas. The documents support a duller account. Publicity for corporations existed before 1917. The state built, for two years and four months, a machine larger than any private office, staffed it with the existing trades of advertising and journalism, and then abolished it. Men who had worked the machine found clients. Some of those men, Bernays among them, wrote books in which the machine became a method and the method became a profession. The books are real. The profession is real. The claim that the profession began in the mind of one nephew of Freud, in a committee he may or may not have chaired, is a product of the profession.

Annotation. What the Creel Committee demonstrates, without any help from a geometric proof, is the material character of propaganda as this treatise has defined it. Definition VI names four conditions: an organised body, a collective object, a multitude, and repetition. The CPI was an organised body created by executive order. Its object was the conduct of a nation at war. Its multitude was the domestic population and the foreign audiences of Compub. Its repetition was the Four Minute Men, the news service, the artists, and the advertisers, working daily until the Armistice stopped the domestic work and the following summer stopped the foreign work. Remove any of the four conditions and the remaining three do not constitute the apparatus that Record Group 63 describes. A clever paragraph, spoken once, by a man without an office, is not the CPI.

The same documents illustrate Definition VII, which names coercion as influence that operates by physical constraint. The committee's work on morale and news sat beside conscription, espionage statutes, and the ordinary violence of a war. Opinion work in 1917 to 1919 was never the sole instrument. It was the cheaper instrument, used where force was already in place or promised. A commentary that treated the CPI as a triumph of technique over matter would have to ignore the Army. The documents do not ignore the Army. They list the Secretaries of War and Navy as members of the committee.

Definition IX names interest as the direction of an agent's endeavour as determined by its actual conditions. A man who has served a wartime information apparatus, and who then opens an office that sells the methods of that apparatus to corporations, has an interest in the belief that those methods are transferable, potent, and scarce. His account of what the committee proved is evidence about his situation. It is weaker evidence about what the committee proved. If Propositions VII and VIII of this treatise hold, access is a necessary part of the definite cause and the seller's account is biased. The Creel years illustrate the first clause by their scale and the second by the career that followed them. They do not, by illustrating a definition, establish that any later commercial campaign worked.

The soldier's question about 1917 is not whether Creel was wicked. The question is what the United States actually built, how long it lasted, who staffed it, and what can be shown about one junior figure's rank. The answers are: an independent committee, two years and four months, advertisers and journalists and seventy-five thousand timed speakers, and a chairmanship attested by Cull and absent from the National Archives' own administrative list. That is enough to begin. It is not enough to found an industry on a name.

Chapter 23

Crystallizing Public Opinion and Propaganda

Annotation. In 1923 Boni and Liveright published Crystallizing Public Opinion. [H] It is the book in which Bernays named the trade. Its first chapter opens by announcing that a new phrase has come into the language: counsel on public relations. The phrase is the merchandise. A press agent is a man you hire to get a notice into a paper. A counsel on public relations is a man you hire as you hire a lawyer, at a fee that implies a profession, a body of knowledge, and a place beside the directors. The book exists to make that implication stick. Bernays then quotes Walter Lippmann at length, including a sentence of Bernays's own wording in which the significant revolution of modern times is not industrial or economic or political, but the revolution which is taking place in the art of creating consent among the governed. Lippmann had written that a revolution was taking place, infinitely more significant than any shifting of economic power. The wording that circulates as Lippmann's is Bernays's. The appropriation is not an acknowledgement of discipleship. The later literature would spend decades treating Bernays as the source of ideas that, in his first professional book, he was still busy lifting from a man who had published them the year before.

Annotation. The phrase "invisible government" does not appear in Crystallizing Public Opinion. [H] It appears in Propaganda, published by Horace Liveright in November 1928, a book of one hundred and fifty-nine pages. The most quoted passage opens Chapter I, "Organizing Chaos," on page 9. The fragment that circulates in isolation is the sentence about an invisible government which is the true ruling power of our country. Isolated, the sentence reads as a confession of conspiracy. The sentences that follow it reverse that reading. The full opening, checked against the text of Propaganda as it stands on disk, is this:

The conscious and intelligent manipulation of the organized habits and opinions of the masses is an important element in democratic society. Those who manipulate this unseen mechanism of society constitute an invisible government which is the true ruling power of our country. We are governed, our minds are molded, our tastes formed, our ideas suggested, largely by men we have never heard of. This is a logical result of the way in which our democratic society is organized. Vast numbers of human beings must cooperate in this manner if they are to live together as a smoothly functioning society.

The fourth and fifth sentences are the ones the industry prefers not to print. They recast the arrangement as a functional necessity of mass democracy, not as a plot against a public that would otherwise govern itself. Bernays continues at once: our invisible governors are, in many cases, unaware of the identity of their fellow members in the inner cabinet. The inner cabinet does not meet. Its members do not know one another. The image is not a cabal. It is a division of labour so ordinary that the labourers need not be introduced. A reader who stops after the second sentence has been sold a thriller. A reader who continues to the fifth has been sold a justification. The book is in the business of selling both, to different buyers, with the same pages.

Annotation. The hedge is more precise than the famous fragment, and it is the thesis of this treatise in the author's own words. In Chapter IV, "The Psychology of Public Relations," Bernays names his sources. [H] Trotter and Le Bon, he writes, approached the subject in a scientific manner. Graham Wallas, Walter Lippmann, and others continued with searching studies of the group mind. Then he asks whether, if we understand the mechanism and motives of the group mind, it is not possible to control and regiment the masses according to our will without their knowing it. The next sentence is the one the trade dropped: "The recent practice of propaganda has proved that it is possible, at least up to a certain point and within certain limits." The limit is in the founding document. The industry inherited the claim and dropped the limit.

The sentences that follow the hedge are not much more boastful than the hedge itself. Mass psychology, Bernays writes, is as yet far from being an exact science, and the mysteries of human motivation are by no means all revealed. Theory and practice have combined with sufficient success to permit the practitioner to know that in certain cases some change in public opinion can be effected with a fair degree of accuracy by operating a certain mechanism, as a motorist regulates speed by the flow of gasoline. Scientific accuracy of results is not to be expected, because many of the elements of the situation must always be beyond his control. This is not the language of omnipotence. It is the language of a man offering a service for sale and taking care not to promise a result he could be sued for missing. The later legend forgot the care. It remembered the gasoline.

Annotation. The two dates attached to The Engineering of Consent are frequently conflated, and the conflation is a small instance of the larger habit of treating Bernays's titles as events. [H] He published an article of that title in The Annals of the American Academy of Political and Social Science, volume 250, March 1947, pages 113 to 120. He later edited a book of the same title for the University of Oklahoma Press in 1955. Britannica records that he edited The Engineering of Consent (1955). The 1955 volume is a multi-author work, not a single-authored treatise. An article in 1947 and an edited collection in 1955 are two objects. Writers who need a founding moment collapse them into one, as they collapse the CPI into a single inventor and Propaganda into a single sentence. The collapse is the method. Precision about titles is not pedantry. It is the refusal to let a brand stand in for a bibliography.

Annotation. Biography of an Idea, published in 1965, is his own account of his career, and the source most often used, and most often doubted, by later writers. [H] The New York Times obituary of 1995 treats it as the standard narrative. Tye, who read the papers behind the narrative, describes the book as devoted to everywhere that Eddie Bernays had been quoted, written for an audience of one. Self-mythology is not a metaphor here. It is a description of a text whose citations are to its author's press. A memoir of that kind is a primary source for what the author wished to have believed. It is a secondary source, and a contaminated one, for what happened. The contamination is structural. Definition IX already names the reason: the interest of an agent is the direction of its endeavour as determined by its actual conditions, and a man whose actual conditions include the sale of his own centrality will produce an account in which he is central. The account is evidence of the sale. It is the weakest evidence of the centrality.

The books, taken together, perform a single operation. Crystallizing Public Opinion announces a profession. Propaganda supplies the profession with a rhetoric of necessity and a rhetoric of power, joined by a hedge. The 1947 article and the 1955 collection offer a cleaner name for the same offer. The 1965 memoir converts the offer into a life. None of these texts is a measurement. None of them is an archive of effects. They are what a counsel on public relations would be expected to produce if his first and lasting client were himself. The reader who treats them as the history of the campaigns they mention has already accepted the terms of the retainer.

What the documents illustrate about Definition VI is again the four conditions, now stated as a commercial prospectus rather than as an executive order. Bernays's "invisible government" is an organised body only in the loose sense of a division of labour. Its collective object is the smooth functioning of a mass society. Its multitude is the hundred and twenty million of 1928. Its repetition is implied by the mechanism itself. The illustration is literary. It does not establish that such a body existed as he described it, and the sentence about fellow members who do not know one another is already a withdrawal of the strongest reading. What the documents illustrate about Definition IX is plainer. A book that tells a corporation it cannot live without a new kind of counsel is a book whose interest is the fee. Propositions VII and VIII, if they hold, say that access is a necessary part of the definite cause and that the seller's account is biased. Propaganda is the seller's account. The distribution, in 1928, was Horace Liveright's, and a hundred and fifty-nine pages.

Chapter 24

The Campaigns, Separated from the Memoir

The memoir treats the campaigns as proofs. The record treats them as episodes, some attested, some inflated, some attributed to Bernays by a later habit of putting his name on other men's work. This section separates the two. Each campaign is stated first as what the documents support, then as what the legend added, and the addition is marked as a correction where the brief and the dossier require a correction. No campaign is used to prove a proposition. A campaign that is well documented may still have been ineffective. A campaign that is poorly documented may still have happened. The two questions are not one question.

Annotation. Torches of Freedom, Easter Sunday 1929. Working for the American Tobacco Company, makers of Lucky Strike, Bernays consulted the psychoanalyst A. A. Brill and then staged an event in which fashionable young women lit cigarettes during the Easter parade on Fifth Avenue. [H] Tye recounts the consultation and the staging. The New York Times obituary of 1995 repeats the standard account: he sponsored demonstrations in which debutantes gathered on street corners to light up, and the cigarettes were even called torches of freedom. The event happened. That much is not in serious dispute. The scale of the effect is.

The correction is this. Vanessa Murphree, writing in American Journalism in 2015, examined more than two hundred and fifty newspaper articles and concluded that the idea that the news media fell for the event in a significant way is a Bernays-driven myth. The coverage indicates that the impact of the event was likely never as extensive or persuasive as Bernays and some have suggested. Historians had long relied on Bernays as their only source of information. Several contemporary clippings dismissed the march as insignificant. Murphree's later teaching note on the same episode criticises documentary treatments for using stock imagery and ignoring those contemporary dismissals. The correction is not that nothing occurred on Fifth Avenue at Easter in 1929. The correction is that the occurrence was converted, by its organiser and by writers who had no other source, into a founding miracle of the trade. A miracle that has one witness, and that witness is the man who billed for it, is not a measured effect.

The wording of Bernays's own later claims about the march was not checked, for this dossier, against the primary scrapbooks at the Library of Congress, which remain under copyright restriction. The substance, a staged event and an exaggerated effect, is supported by peer-reviewed scholarship. The memoir's particular sentences about how completely the nation changed its mind that Sunday are not verified here. They belong with the other unverified items listed at the end of this Part.

Annotation. The green Lucky Strike pack. When surveys indicated that women disliked the forest-green pack because it clashed with their clothes, Bernays did not change the pack. [H] He set out to make green fashionable, writing to designers, department stores, and prominent women, and organising luncheons, balls, exhibitions, and green window displays. The Encyclopedia of World Biography records the method. The New York Times obituary calls the campaign a brilliant success, according to sales figures. That sentence is the newspaper's claim, not a measured effect established in this Part. The obituary does not reproduce the figures, the interval, or the comparison. What is attested is the work: the letters, the luncheons, the displays, the refusal to alter the pack. What is not attested, at the level of evidence this treatise will accept, is the conversion of those efforts into a known increment of sales. A newspaper's praise of sales figures is a kind of evidence about the newspaper and about the legend. It is a weaker kind of evidence about the market.

Annotation. "Reach for a Lucky instead of a sweet." This attribution is a correction, and it must be stated as a correction rather than as lore. [H] The slogan belonged to the American Tobacco Company and its management, notably president George Washington Hill. It is documented as a company advertising campaign in the Journal of Historical Research in Marketing (2010) and in the surviving advertisements themselves, including those held in the Stanford Tobacco Advertising archive. No source examined for the dossier establishes a Bernays role in the sweets-versus-slimming angle. The habit of attaching the slogan to him is an instance of the larger habit Tye describes: because Bernays genuinely did shape some large things, audiences assume he did everything he claimed, and then assume he did things he did not even claim. The sweets campaign had a corporation, a president, a budget, and a media buy. Those are the causes the documents name. The technique, if a technique is wanted, was merchandise. The merchandise was Hill's.

Annotation. Coolidge, 1924. To counter Calvin Coolidge's stiff public image during the 1924 election, Bernays organised White House entertainment featuring Al Jolson and other Broadway performers. [H] Chernow independently records that Bernays tried to brighten Coolidge's dour image during the 1924 campaign by importing Al Jolson and forty other Broadway performers to the White House. Richard Gunderman, writing in The Conversation in 2015, records pancake breakfasts and White House concerts featuring Jolson and other Broadway performers. The presence of Broadway performers, including Jolson, at the White House in connection with the 1924 campaign is therefore attested by two writers who are not copying a single sentence from each other. The Library of Congress holds Bernays's own typed note, "Breakfast with Coolidge," dated 8 February 1962, a retrospective document, not a 1924 diary.

The correction is this. The date and the "pancake breakfast" form are contested. A common internet rendering has it that on 17 October 1924 Bernays engineered a pancake breakfast with vaudevillians for Coolidge. The dossier could not load the body of the Library of Congress typescript, blocked by a Cloudflare interstitial, and could not load the Evening Star of 17 October 1924 to check the date or the word "pancake." Gunderman and Chernow confirm Broadway performers at the White House. They do not, in the pages used here, fix the date or certify the breakfast. A second correction follows. The claim that, with Bernays's help, Coolidge won the 1924 election is unsupported opinion. Gunderman states a causal relation that the electoral record does not isolate. Coolidge was the incumbent. The opposition was divided. The economy of 1924 is not a public relations account. To credit a morning of actors with a national result is to practise the profession one is describing. This Part does not practise it.

Annotation. Futurama, 1939 New York World's Fair. This attribution is a correction. [H] Futurama was the General Motors exhibit designed by Norman Bel Geddes, a ride presenting a model of America in 1960 and widely credited with introducing the expressway idea to a mass audience. Bel Geddes published Magic Motorways in 1940. Bernays's documented connection to the fair is as its publicity director. The Encyclopedia of World Biography states the appointment: in 1939 he was the publicity director for the New York World's Fair. A direct Bernays role in Futurama itself was not verified. The correction is required because the legend, once it has a publicity director and a famous exhibit in the same year, will put the director inside the exhibit. The documents put Bel Geddes inside the exhibit and Bernays in the fair's publicity office. Those are different rooms. Distribution of attention to the fair is not design of a General Motors ride. The first may have been Bernays's work. The second, on the evidence examined, was not.

Annotation. Procter and Gamble, Ivory soap. The soap-sculpting contests are attested. [H] Bernays hired a medical consultant to survey hospitals on their preference for white, unperfumed soap, giving Ivory a medical endorsement, and established the national soap-sculpting contests judged by artists that, in the language of the Encyclopedia of World Biography and the New York Times obituary, eventually involved millions of schoolchildren. The contests are a matter of record in the same sense that a sponsored prize is a matter of record: there were rules, judges, and a corporate sponsor. The figure of millions is the sources' figure, not a count performed here. What the episode shows without that figure is the method at its most ordinary. A company already possessed a product, a colour, a distribution system, and access to schools. The counsel attached a contest and a medical survey to what the company already possessed. The hospital survey did not create a preference for white, unperfumed soap. It collected a preference that hospitals, on the consultant's asking, were prepared to state, and it returned that preference to the company as an endorsement. The sculpture contest did not create a taste for Ivory. It gave schoolchildren a task whose material was the product, and it gave artists a jury whose prestige attached to the task. The power was the access to children and to hospital purchasing. The contest was the packaging of that access. A firm that can enter the schoolroom does not need a theory of the crowd. It needs a superintendent.

Annotation. United Fruit and Guatemala. This is the campaign for which the documentary record is strongest, and it is the campaign in which the causes named by this treatise are most visible. [H] Bernays began working for the United Fruit Company in the early 1940s under Samuel Zemurray, first to sell more bananas and then, as Guatemalan politics shifted, to wage a propaganda campaign against the government of Jacobo Arbenz. Using fifty-three boxes of Bernays's United Fruit papers released by the Library of Congress after his death, Tye documents a programme that included placing articles in major magazines, staging press tours for publishers, and building front groups such as the Middle America Information Bureau. Thomas McCann, a United Fruit public relations official, disputed Bernays's claim that journalists were free to report what they found. What the press would hear and see, McCann said, was carefully staged and regulated by the host. Chernow summarises the outcome: Bernays circulated unflattering information to influential American newspapers, stigmatising Arbenz with the Communist label and softening up public opinion for the eventual CIA-sponsored overthrow of the reform-minded government.

The correction here is not that the work has been exaggerated in the manner of Torches of Freedom. The correction is of a different kind. This is the case where distribution plus state force, not technique, did the work. The fifty-three boxes show access to magazines, to publishers, and to the itinerary of a press tour. They show a named front, the Middle America Information Bureau, which is the institutional form of a voice that does not wish to be heard as United Fruit's. McCann's testimony shows that the tour was a regulated environment. Chernow's summary shows the Communist label and the CIA. The sequence of the retainer itself is evidence about interest: first bananas, under Zemurray, then, as Guatemalan politics shifted, a campaign against a government. The shift was not a discovery about the psyche of the American reader. It was a change in the client's political situation, and the office moved with the client.

The overthrow of Arbenz was not a public relations effect. It was a covert operation by a state, prepared in an atmosphere that a corporate publicity apparatus had helped to thicken. A man who can place a story in a newspaper is not a man who can remove a government. The government was removed by the agency that removes governments. The stories were the cheaper instrument, used where force was already in prospect. Definition VII is not a decoration on this paragraph. It is the name of the cause that the legend of technique is designed to hide. The Communist label is Definition VI in miniature: an organised body, a collective object (the legitimacy of a foreign government), a multitude (the American newspaper public), and repetition (articles, tours, a bureau). The label did not have to be a syllogism. It had to be a word already stocked in the reader's imagination, attached to a name the reader did not know, by channels the reader did not own. Lippmann had already described that attachment in 1922. Bernays, in the years of the United Fruit retainer, had the magazines.

The smoking papers belong beside the banana papers, not as a campaign with a name, but as a measure of the narrator. Tye records that Bernays knew of early studies linking smoking and cancer while publicly promoting cigarettes, and later claimed ignorance, a claim his own papers contradict. [H] The contradiction is not offered here as a moral discovery. It is offered as a documentary one. The same archive that supports his centrality in the United Fruit account undermines his later account of what he knew about tobacco. An unreliable narrator is not unreliable in patches, reliable where the story is flattering and unreliable only where it is not. He is the same man in both files.

Annotation. Government war work in the Second World War is recorded in outline. [H] The Encyclopedia of World Biography states that Bernays's services were called upon by the Army, the Navy, and the Commerce and Treasury departments, and that after the war he was involved in the government's foreign information programme. The outline is thin. It establishes that the state remained a client. It does not establish a second Creel Committee with Bernays as Creel. The temptation to write that second committee into the record is the same temptation that writes him into Futurama and into Hill's slogan. This Part declines it.

Taken as a set, the campaigns do not support the memoir's portrait of a man whose technique moved nations. They support a narrower portrait. Some events were staged and later inflated (the Easter parade). Some advertising belonged to the client and was later hung on the counsel (the sweets slogan). Some famous objects in the same year were not his (Futurama). Some political theatre is attested and its electoral consequence is not (Coolidge). Some programmes are attested as work and praised as sales by a newspaper (the green pack). Some contests happened (Ivory). One campaign is well documented and its decisive cause was a state (United Fruit). The pattern is not mysterious. It is what one would expect if the power were distribution, access, and capital, and if the man who sold the power were also the man who wrote the history of the sale.

The documents illustrate Definition VI wherever an organised body addressed a multitude on a collective object with repetition: a tobacco company, a soap company, a fruit company, a fair, a White House. They illustrate Definition VII where the CIA enters the Guatemalan account. They illustrate Definition IX in every paragraph of the memoir that treats a billing as a cause. They illustrate Propositions VII and VIII, if those propositions hold, without proving them: the power that can be shown is access to publishers, to schools, to a fair's publicity office, to a press tour, and to a state. The seller's account is biased evidence of what that access accomplished. None of this is a demonstration that 1929 changed the smoking habits of American women. Murphree's two hundred and fifty articles are the relevant court for that claim, and they returned a thinner verdict than the memoir.

Chapter 25

Doris Fleischman

Annotation. After the First World War Bernays and Doris Fleischman, a journalist he had hired in 1919, opened their own public relations office. [H] Early clients included the United States War Department and the Lithuanian government. They married in 1922. Tye records the wedding in the New York Municipal Building and the subsequent registration of the bride at the Waldorf-Astoria under her maiden name, an episode reported in more than two hundred and fifty newspapers. The day of the marriage is not settled in the sources used here. Tye places the matrimonial item on the evening of 16 September. Other sources give the year only. The unsettled day is listed among the unverified items at the end of this Part. The partnership is not unsettled.

Annotation. Fleischman was not an assistant. She was a co-architect of the practice. The entry on her in the Jewish Women's Archive, written by Bernays's own daughter Anne Bernays, states the division of labour without romance. [H] Nominally his partner, she wrote most of the firm's press releases, speeches, and important letters. She was the first married American woman issued a passport in her own name, in 1925. She worked with Bernays for sixty-one years. Susan Henry's study, available as ERIC document ED298562 (1988), documents that they maintained a partnership in one of the country's premier public relations firms, while little notice was given to Fleischman despite her high visibility. Tye is blunt about the asymmetry. Her role was absolutely essential. Eddie Bernays did a whole lot to promote Doris's ideas but not to promote Doris.

The legend writes her out. That deletion is not an accident of space and not a consequence of the evidence. It is part of the self-mythology, which requires a single inventor, as the CPI story requires a single ancestor and the Torches story requires a single Sunday. A partnership that lasts sixty-one years, in which one partner writes most of the releases, the speeches, and the important letters, is not a firm with a father. It is a firm with a division of labour that the father's books do not describe. Anne Bernays's testimony is the least convenient kind. It comes from inside the woman, from a daughter who cannot be dismissed as a rival practitioner, and it names the writing, which is the actual product the firm sold. Press releases, speeches, and important letters are the merchandise. The man who did not write most of them remains the name on the merchandise.

Annotation. The firm's power, so far as the documents speak to it, rested on access to capital and to media owners rather than on a secret of technique. [H] By the 1940s the client roster was a map of American corporate and media power: General Motors, General Electric, Procter and Gamble, Time Inc. CBS, and NBC, alongside United Fruit and the Waldorf-Astoria. The New York Times obituary lists them. Access to publishers and broadcasters was thus structurally embedded in the practice. A counsel who numbers Time, CBS, and NBC among his clients is not standing outside the channels and working on them by cunning. He is inside the channels. Bernays's pricing policy put the firm in the same economic class as its clients. He argued, Tye records, that the only way to get people like the President or the heads of big companies to take the firm seriously was to charge a fee equal to the salary they were making. The fee is not a measure of effect. It is a measure of the class the firm intended to occupy. A high fee is itself a signal, and the signal works on the buyer, who has an interest in believing that what was expensive was potent.

Fleischman's passport of 1925 is sometimes told as a stunt and sometimes as a principle. The documents support both descriptions at once, which is to say they support a practice in which a principle and a stunt were not two things. A woman who insisted on her own name, and a firm that knew what a newspaper would do with that insistence, are not a contradiction. They are a partnership. The legend that followed kept the passport as a colourful item in his biography and dropped the information that she wrote the letters. The item is colourful. The letters are the work.

Definition IX applies here without any strain of commentary. The interest of the named partner was the direction of a career that had been built on being named. The interest of the writing partner, as the documents stand, was not independently converted into a rival legend. Tye's sentence about promoting the ideas and not the woman is a sentence about interest, not about affection. Propositions VII and VIII, if they hold, say that access is a necessary part of the definite cause and that the seller's account is biased. Fleischman wrote for the distribution. Bernays sold the account. The later books are his. Most of the releases were hers. The industry quotes the books.

Chapter 26

The Mythology

Annotation. The strongest scepticism about the Bernays legend comes from the people who read him most carefully, including the biographer who had the run of his papers. Larry Tye's The Father of Spin (1998) is the central work. [H] Tye worked through more than eight hundred boxes of Bernays's papers at the Library of Congress and interviewed more than a hundred people who knew him. Chernow, reviewing the book, records those figures. Tye concludes that Bernays was an unreliable narrator who gave himself a more central role in things than he might have deserved in lots of his campaigns. The explanation of the self-mythology is not mysterious. Because Bernays genuinely did shape some large things, audiences assume he did everything he claimed. The assumption is a habit of inference, and it is the habit on which a career of self-description feeds. The same papers that enlarge some campaigns shrink others. They show a man who knew of early studies linking smoking and cancer while he was still promoting cigarettes, and who later claimed ignorance. Tye states that the claim of ignorance is contradicted by the files. An archive used as a trophy case will not remain a trophy case once a second reader enters it. Eight hundred boxes are not a confession. They are a condition of correction. A legend built on a memoir can survive any number of magazine profiles. It has more trouble surviving a reader who has the incoming correspondence.

Annotation. Chernow's review, "First Among Flacks," published in the New York Times Book Review in 1998, states the puffery without ornament. [H] In no area of his life did Edward L. Bernays apply his powers of puffery more artfully than in fashioning his own image. Tye, Chernow writes, effectively deflates Bernays's pompous claims for himself. The review is not a hatchet piece by a stranger to American business history. It is a judgement by a biographer of Rockefeller and Morgan, delivered in the newspaper that had printed the legend in its obituary three years earlier. The newspaper contained both the myth and the correction. The later popular literature preferred the obituary.

Annotation. Murphree's 2015 article in American Journalism is the model of what a correction looks like when it is done with contemporaneous print rather than with a memoir. [H] She shows that the Torches of Freedom march was a Bernays-driven myth in its supposed impact, and she criticises documentary treatments of the episode for using stock imagery and ignoring contemporary articles that dismissed the march as publicity. The importance of the article for this Part is not only the Easter finding. It is the method. If historians had relied on Bernays as their only source, then the historical literature was, for that episode, an extension of the account. The client, in the last instance, was the same man. Peer review of a memoir is not the same labour as peer review of two hundred and fifty newspapers. Murphree performed the second labour. The myth did not survive it intact.

Annotation. Adam Curtis's 2002 BBC series The Century of the Self is the most widely circulated version of the thesis that Freud's ideas were used to manipulate American consumers and voters. [H] It is a popularising polemic, not a peer-reviewed work, and its own framing confirms the polemic. In episode one, Curtis says that the series is about how those in power have used Freud's theories to try to control the dangerous crowd in an age of mass democracy. The sentence is a thesis statement, not a finding. It organises four hours of archival footage and interview around a causal chain that runs from Vienna to the American supermarket. The chain is memorable. Memorability is not a documentary virtue. It is a cinematic one.

Media Lens, responding in 2002, argued that the series ultimately reproduced exactly the kind of deceptive spin that Bernays used to camouflage the truth about big business control, and that Freud's theories were incidental, a sideshow to elite efforts at popular control, which were driven by material conflict rather than psychoanalytic doctrine. The point stands independently of Media Lens's politics. Bernays's books cite Wilfred Trotter, Gustave Le Bon, Graham Wallas, and Walter Lippmann alongside Freud. A 2013 study by Moloney and colleagues in the International Journal of Communication argues that his key theoretical ideas were lifted from Lippmann and rebranded. The Freud kinship was real. The inference from kinship to a psychoanalytic engine of consumerism is a narrative convenience. It gives the viewer a European villain of ideas and an American villain of technique, joined by blood. The books themselves, which are the relevant documents for what Bernays claimed to be doing, name a British surgeon (Trotter), a French crowd psychologist (Le Bon), a British political scientist (Wallas), and an American journalist who had already used the phrase the industry later treated as Bernays's (Lippmann). Freud is not absent. He is not the engine.

The Curtis series did the industry a service the industry could not have purchased more cheaply. It took the seller's account, added Freud, added footage, and distributed the result to an audience larger than Propaganda ever reached. The series is therefore itself an instance of the thing this Part is describing: a claim about power, carried by a channel, repeated until it became the furniture of educated conversation. To treat The Century of the Self as evidence about 1929 is to confuse a downstream popularisation with an upstream record. The record is Tye's boxes, Murphree's newspapers, McCann's sentence about the staged tour, Chernow's deflation, and the pages of Propaganda in which the hedge still sits where Bernays left it.

The Freud-to-consumerism chain has a further documentary weakness that does not depend on anyone's politics. Bernays was Freud's nephew twice over. That kinship is a fact about a family. It is not a fact about a mechanism. A man may cite his uncle, may visit him, may boast of the relation in a market that honours European science, and still write a professional book whose named authorities are Trotter, Le Bon, Wallas, and Lippmann. The boast and the bibliography are two objects. Curtis organised the series around the boast. Moloney organised a paper around the bibliography. This Part follows the paper, not because papers are virtuous, but because the pages of Propaganda are the pages that exist. A polemic that needs Freud as the engine must treat those pages as incomplete. They are not incomplete. They are inconvenient.

Annotation. The myth has a structure, and the structure is economical. A wartime committee becomes a unique origin. A partnership becomes a father. A staged parade becomes a change in the nation's morals. A client's slogan becomes the counsel's slogan. A fair's publicity office becomes an exhibit designed by someone else. A press tour and a covert operation become a technique. A BBC series becomes a seminar. At each step the name Bernays absorbs a cause that the documents assign elsewhere: to Creel, to Fleischman, to Hill, to Bel Geddes, to the CIA, to Lippmann, to the newspapers that did not, in Murphree's sample, fall for the march. The absorption is the work of interest. Definition IX again names it. A man whose endeavour is determined by the sale of centrality will produce a map on which all roads meet at his office. Eight hundred boxes, read by a reporter who was not on the payroll, redraw the map. The roads still pass near the office. They do not all begin there.

The myth is also a lesson in what this treatise calls, in its opening Part, the abuse of words. The word "father" does causal work it has not earned. It converts a participant into an origin, and an origin into a necessity: if the craft has a father, the craft has a nature, and the nature can be taught as his. Chernow's title, "First Among Flacks," is already a correction of paternity. First among is a rank in a field that existed. Father of is a rank in a field that must be invented in order to be owned. Ivy Lee from 1906, Byoir from the same committee, Fleischman from the same office, Lippmann from the same war and the earlier book: the field is crowded once the documents are allowed to name more than one person. The mythology is the refusal of that crowd.

If Propositions VII and VIII hold, the mythology is what a seller's account looks like when it has been repeated long enough to become a curriculum. Power is distribution: Tye's book distributed a correction through a commercial press and a newspaper review, and still lost, in the popular mind, to a television series. The seller's account is biased: the strongest evidence in this section is the boxes, the newspapers, and the pages that name other authors. The weakest is the story Bernays told about himself, and the stories told by people who had read only him. That ranking is a ranking of sources. It is not a proof that any definition in Part I is true. It is what the sources look like when they are ranked.

Chapter 27

Le Bon, Lippmann, and the Manufacture of Consent

Annotation. Gustave Le Bon's La Psychologie des foules was published in 1895 and in English as The Crowd: A Study of the Popular Mind in 1896. [H] It was his most popular work. Encyclopaedia Britannica summarises its thesis: the conscious personality of the individual in a crowd is submerged and the collective crowd mind dominates; crowd behaviour is unanimous, emotional, and intellectually weak. Britannica also records that the emerging profession drew on the book to persuade kings and business potentates that they were no longer the rulers, and that the crowd, the public, was now sovereign and must be pleased. Bernays cites Trotter and Le Bon by name in Propaganda as men who approached the subject in a scientific manner. The citation is not hidden. It is on the page on which he asks whether the masses can be regimented without their knowing it. The intellectual furniture of Chapter IV is nineteenth-century crowd theory and its early twentieth-century elaborations, not a private discovery of 1928.

Annotation. Walter Lippmann's Public Opinion was published by Harcourt, Brace in 1922. [H] It developed the concepts of the pseudo-environment and the stereotype, and it is the source of the term "manufacture of consent." The phrase appears in Lippmann's own text in Chapter XV, "Leaders and the Rank and File." The passage, checked against the chapter as it stands on disk, is this:

That the manufacture of consent is capable of great refinements no one, I think, denies. The process by which public opinions arise is certainly no less intricate than it has appeared in these pages, and the opportunities for manipulation open to anyone who understands the process are plain enough. The creation of consent is not a new art. It is a very old one which was supposed to have died out with the appearance of democracy. But it has not died out. It has, in fact, improved enormously in technic, because it is now based on analysis rather than on rule of thumb.

Lippmann continues, and the continuation is given here as paraphrase of the same chapter, not as a second block quotation. As a result of psychological research, coupled with the modern means of communication, the practice of democracy has turned a corner. A revolution is taking place, infinitely more significant than any shifting of economic power. Within the life of the generation now in control of affairs, persuasion has become a self-conscious art and a regular organ of popular government. None of us begins to understand the consequences, but it is no daring prophecy to say that the knowledge of how to create consent will alter every political calculation and modify every political premise. Under the impact of propaganda, not necessarily in the sinister meaning of the word alone, the old constants of our thinking have become variables. It is no longer possible, Lippmann writes, to believe the original dogma of democracy, that the knowledge needed for the management of human affairs comes up spontaneously from the human heart.

The manufacture paragraph just quoted is Lippmann's, Chapter XV, 1922. It is not the sentence Bernays printed in Crystallizing Public Opinion. Crystallizing invents its own revolution sentence and garbles the persuasion and dogma sentences; it does not quote the manufacture paragraph. The later habit of treating "manufacture of consent" as a Bernays coinage, or as a later radical discovery about the press, is a habit that cannot survive contact with Chapter XV. The later habit of treating Bernays's revolution sentence as Lippmann's is a habit that cannot survive contact with Crystallizing.

Annotation. The claim that John Dewey coined the phrase is a misattribution. [H] The phrase is in Lippmann, Chapter XV, in the sentences just quoted. No source located for the dossier shows Dewey coining it. Dewey did review Public Opinion in The New Republic on 3 May 1922, and he later published The Public and Its Problems (1927), in which he rejected Lippmann's technocratic solution. The text of the Dewey review was not accessible for verification. This Part therefore treats Dewey's wording as unconfirmed. It does not treat Dewey's authorship of the phrase as an open question. Authorship of the phrase, on the evidence of the book itself, belongs to the book. A review of the book cannot be the origin of a sentence printed in the book under review. The misattribution persists because a debate is easier to remember than a chapter number. The Lippmann-Dewey debate is a convenient classroom object. Chapter XV is a set of paragraphs. The paragraphs have the phrase. The classroom object has the fame.

Annotation. The Phantom Public (1925) argued that the public is a phantom that can only intermittently intervene, and it contains the line that the public must be put in its place […] so that each of us may live free of the trampling and the roar of a bewildered herd. [H] The "bewildered herd" is Lippmann's. It is not Bernays's. It belongs to a book that appeared three years after Public Opinion and three years before Propaganda. The sequence of publications is itself an argument against the legend of the inventor. Le Bon in 1895 and 1896, Lippmann in 1922 and 1925, Bernays in 1923 and 1928: the dates do not support a transmission from Bernays to the others. They support a transmission toward him.

Stuart Ewen, in PR!: A Social History of Spin (1996), captures the relationship in a sentence that the later literature has found useful: what Lippmann set out in grand, overview terms, Bernays is running through in how-to-do-it terms. Wikipedia, drawing on published scholarship, describes Lippmann as Bernays's unacknowledged American mentor. Unacknowledged is the correct adjective for the relationship as it appears in Propaganda, where Lippmann is named in a list and not as a source of the book's central phrases. Crystallizing Public Opinion is more open, because it quotes. Propaganda is more famous, because it boasts. The boast, once separated from the quotation marks, became a founding.

Annotation. The wartime link is direct on both sides. Lippmann, like Bernays, served the Wilson propaganda effort. [H] Cull records the service. The two men were not, in 1917 to 1919, a theorist and a pupil working in different worlds. They were participants in the same state's information work, one of them already a journalist of standing, the other a theatrical publicist whose exact rank on the committee remains, as already noted, a one-source attribution. After the war Lippmann wrote the book that named the manufacture of consent as an old art improved by analysis. Bernays wrote the book that named an invisible government and offered to represent it for a fee. The difference is the difference between a diagnosis and a prospectus. Moloney and colleagues (2013) describe the second as a rebranding of the first. The description fits the pages. It does not require a conspiracy of silence. It requires a professional interest in being the man who is cited, which is the interest Definition IX already names.

Le Bon supplied a psychology of the crowd that flattered the practitioner by making the crowd unanimous, emotional, and weak. Lippmann supplied an account of democracy in which symbols, stereotypes, and the manufacture of consent were already the ordinary tools of leadership, and in which the newspapers were already organs of that manufacture. Bernays supplied a title, a fee structure, and a set of anecdotes. The intellectual lineage is not a descent from a master of technique. It is an ascent of a salesman through other men's analysis. The analysis remains more careful than the sales. Lippmann's Chapter XV is a description of leaders, symbols, morale, casualty figures, and the official's control over access to facts. It is not a promise that a counsel can regiment the masses at will. Bernays's Chapter IV asks that question and then, in the next sentence, inserts the limit. The industry, repeating Bernays, dropped Lippmann's patience and Bernays's limit together.

What these pages illustrate about Definition VI is that the word propaganda, in Lippmann as in this treatise, names an organised practice with a collective object, not a sin. Lippmann writes of propaganda not necessarily in the sinister meaning of the word alone. The reservation is the same reservation this book's Definition VI makes by remaining silent on truth. What they illustrate about Definition IX is the residual relation Ewen names: the how-to is the interest, the overview is the source. What they illustrate about Propositions VII and VIII, if those propositions hold, is again distribution and the weakness of the seller's account. Lippmann's distribution in 1922 was Harcourt, Brace and the authority of a man who had already been read. Bernays's distribution in 1928 was Liveright and the authority of a man who had already been hired. The later curriculum assigned the phrase to the hirer. The chapter on disk assigns it to Lippmann.

Chapter 28

How Public Relations Replaced Propaganda

Annotation. "Propaganda" had no negative charge for Bernays in 1928. He used it in his title and defined it as necessary. The word then curdled, and its replacement was a deliberate act of professional branding. [H] Gunderman states it plainly: propaganda had acquired a somewhat pejorative connotation, which would be further magnified during the Second World War, so Bernays promoted the term public relations. He had been building the alternative vocabulary since Crystallizing Public Opinion (1923), which opened by announcing that a new phrase had come into the language, counsel on public relations. He coined and publicised the title. He lectured on the subject at New York University in 1923, the first such lectures at a major university. In his last years he pressed for practitioner licensing. The Encyclopedia of World Biography and the New York Times obituary record the lectures and the late campaign for a licence. The rebranding is therefore not a post-war improvisation. It is contemporaneous with the first book. The 1928 title Propaganda is, in that light, a residual boldness, a word he was still willing to print while the office had already begun to sell the cleaner name. No letterhead sequence establishing the substitution has been found for this treatise, and none has been invented.

Annotation. The professionalisation was real and partial. Public relations became a distinct occupation with its own vocabulary, university teaching, and fee structures. [H] Bernays worked to make practitioners salaried like executives rather than retained like press agents. Tye records the fee argument in that form. A salary is a claim to belong inside the firm. A retainer is a claim to be called when needed. The difference is status, and status is one of the things a counsel on public relations sells. Yet acceptance of the profession's self-description was eroded by fascism. Curt Olsen, in Public Relations Tactics (2005), argues that the public largely accepted Bernays's sunny view of propaganda, an acceptance eroded by fascism in the Second World War era, and that Bernays's language let softer terms such as education replace harder ones such as indoctrination. The substitution of words is the same operation as the substitution of public relations for propaganda, performed one register lower. Education, public relations, counsel: each is cleaner than the word it displaces. Each displaces a word that had begun to name a cause the speaker no longer wished to have named.

Annotation. Public mistrust of official government presence in the media is itself traceable to the CPI. [H] Congress wound the committee down with visible haste in 1919. Cull notes that propaganda became, and remains, one of the dirty words of American politics. The dirt is historical. It is not an essence of the activity. The activity continued under the new name, in universities, in firms, in the Army and Navy and Treasury accounts of the next war, and in the foreign information programme afterwards. The dirty word and the clean word named, at the level of mechanism, the same four conditions Definition VI already named: organisation, a collective object, a multitude, repetition. What changed was the willingness of a paying public to hear the old word without flinching. The flinch is an affect. The occupation is a roster of clients.

Annotation. Bernays was sour about the field he had named. In a 1991 interview, quoted in the New York Times obituary of 1995, he said: "Public relations today is horrible. Any dope, any nitwit, any idiot can call him or herself a public relations practitioner." [H] The complaint is the complaint of a man who had spent the last years of a long life pressing for a licence and had not obtained a closed shop. If anyone can use the title, the title cannot support the fee that was supposed to equal a president's salary. The 1991 sentence is therefore consistent with the 1923 sentence. Both are about the phrase. The first announces that the phrase has come into the language. The last laments that the language has not been fenced. Between them sit the books, the campaigns, the partnership, the boxes, and the war that made the old word dirty. The confession in the title of this Part is not a deathbed recantation of wickedness. It is the 1928 hedge, the 1991 disgust, and the long labour of putting a clean name on an apparatus whose power was never the name.

The documents illustrate Definition VI by showing the word propaganda under two evaluations and one mechanism. They illustrate Definition IX by showing a professional interest in the evaluation, not in the mechanism. A dirty word is bad for fees. A clean word, once it is no longer scarce, is also bad for fees. The endeavour, determined by those conditions, produces first a rebranding and then a lament. Propositions VII and VIII, if they hold, add nothing historical to this sequence. They add a reason to look at who owned the channels while the words were being changed. The channels, in 1923, included a major university lecture room and a New York publisher. The channels, in 1991, included the newspaper that would print the lament in the obituary. The technique, in both years, was the phrase itself.

Chapter 29

What the Documentary Record Cannot Settle

This section is a feature of the inquiry, not a failure of it. A documentary Part that converted every gap into a fact would be doing the work the memoir did. The dossier on which these annotations rest ends with twelve items that could not be verified. They are stated here as gaps. They are not filled. They are not used as the basis of a narrative. A later researcher with better access may close some of them. Until then they remain what they are: places where the legend is more confident than the file.

First. Dewey's review as the origin of "manufacture of consent." The phrase is documented in Lippmann's Public Opinion (1922), Chapter XV. The compiler of the dossier could not access Dewey's review in The New Republic of 3 May 1922 to check whether Dewey used the phrase. The common attribution to Dewey appears to be a misattribution. The Dewey text itself was not verified. This Part has treated the misattribution as a misattribution on the strength of Lippmann's pages, and Dewey's wording as unconfirmed.

Second. Bernays's role in "Reach for a Lucky instead of a sweet." The slogan is documented as the American Tobacco Company's own advertising, associated in the historical literature with George Washington Hill. No source examined establishes what Bernays contributed to the sweets-versus-slimming angle. The attribution of the slogan to him is a correction waiting to be unlearned, not a fact waiting to be confirmed.

Third. Any direct Bernays involvement in Futurama. Futurama is documented as Norman Bel Geddes's design for General Motors. Bernays's documented link to the 1939 fair is as its publicity director. Direct involvement in the exhibit was not verified. The two facts may stand beside each other indefinitely without becoming a third fact.

Fourth. The date and form of the 1924 Coolidge episode. Broadway performers, including Al Jolson, at the White House are attested by Chernow and Gunderman. The date 17 October 1924 and the word "pancake" were not confirmed against the Evening Star of that day or against the body of the Library of Congress typescript "Breakfast with Coolidge," which the compiler could not load. The episode happened in some form. The breakfast, as a dated and named object, remains contested.

Fifth. The claim that, with Bernays's help, Coolidge won the 1924 election. This causal sentence, in Gunderman, is unsupported and is treated in this Part as opinion. An election has many causes. The documents used here do not isolate a public relations account among them.

Sixth. The year of publication of Public Relations. Wikipedia lists 1945. Britannica and the Encyclopedia of World Biography list 1952. This Part has not adjudicated the discrepancy and has not used the book as a load-bearing source. A title whose year cannot be stated from concordant references is not a foundation.

Seventh. The title "chaired the CPI Export Service." The title rests on one source, Cull, writing for PBS. National Archives records do not name Bernays in the administrative history of Record Group 63. The chairmanship is therefore a sourced claim, not an archival appointment.

Eighth. The year 1925 as the date of Bernays's American Tobacco work. Some brief accounts refer to 1925 American Tobacco work. The verified sources place his American Tobacco and Lucky Strike campaigns in the late 1920s, with Torches of Freedom in 1929. The earlier year is not established in the dossier.

Ninth. The firm's name changes. No source located in the dossier documents a sequence of names for the Bernays practice. Sources call it his office or firm, established in 1919. A history of letterheads would be a useful article. It is not available here, and it has not been invented.

Tenth. Goebbels as a student of Bernays. Gunderman and Chernow both repeat the claim, drawing on Bernays's autobiography. It has not been verified against German-language sources or against Goebbels's diaries. An autobiography that is already, on Tye's showing, an unreliable narrator's book is not a sufficient source for a transmission into the propaganda ministry of the Third Reich. The item is famous. Fame is not verification.

Eleventh. The wording of Bernays's own later claims about Torches of Freedom. The substance, a staged event and an exaggerated effect, is supported by Murphree's peer-reviewed work. The memoir's and the interviews' particular wording was not checked against the primary documents, which are held in copyright-restricted scrapbooks at the Library of Congress. This Part has therefore not quoted those later claims as if they had been read.

Twelfth. The day of the 1922 marriage. Tye places the matrimonial item on the evening of 16 September. Other sources give the year only. The year is enough for the partnership. The day is not settled.

These twelve gaps have a common shape. In each of them a later writer, or Bernays himself, supplied a precision that the independent record does not, in the sources examined, sustain: a coinage, a slogan, an exhibit, a breakfast, an election, a year, a title, another year, a sequence of names, a European pupil, a quotation, a calendar day. Precision is one of the things the trade sells. It is also one of the things the trade performs on its own past. The correct response is not to replace one precision with another invented in the same style. The correct response is to leave the gap visible, so that the next reader can see where the file stops.

The twelve items also illustrate, as commentary and not as proof, the two propositions this Part has been permitted to invoke. If power is distribution, then several of the gaps are gaps in distribution: a review that could not be fetched, a typescript behind a Cloudflare wall, scrapbooks closed by copyright, German diaries not consulted. The limits of this dossier are material limits. If the seller's account is biased, then the items that rest on the autobiography (the pupil in Berlin, the later wording of the Easter triumph, the centrality implied by every missing date filled in with a flourish) are exactly the items that should have been weakest, and they are. A documentary Part that closed those items by trusting the seller would have betrayed its own court.

What, then, can be said, at the end of the annotations, without converting the record into a theorem? The United States built, in 1917 to 1919, a large official propaganda apparatus and abolished it. Ivy Lee had already been at the work for a decade. Bernays passed through the apparatus in a role that one historian names and the National Archives' administrative history does not. He and Doris Fleischman opened an office. She wrote most of the releases, the speeches, and the important letters. He wrote the books in which the office became a philosophy of democracy. The books quote Lippmann and cite Le Bon, Trotter, and Wallas. They contain a hedge, "at least up to a certain point and within certain limits," which the later industry did not inherit. Some campaigns are attested as work. Some famous objects in his vicinity were other men's work. One campaign, United Fruit in Guatemala, is well documented, and the decisive force in that campaign was a state. Tye read eight hundred boxes and found an unreliable narrator. Murphree read two hundred and fifty newspapers and found a myth. Curtis made a series that made the myth international. The word propaganda curdled. The word public relations replaced it. In 1991 the old man said the new word had become horrible.

That is the documentary residue. It is smaller than the legend and larger than a debunking. A man sold his own power. Where the power existed it was distribution, access to publishers, state training, and capital. The technique was the thing he needed the buyer to want. The confession is not that he was wicked. The confession is that he said as much himself, in the sentence about limits, and that the sentence was not the one that was remembered.

The geometric work of this treatise continues elsewhere. Definitions VI, VII, and IX were not proved by 1917, and 1917 was not proved by them. What the years 1917 to 1995 supply is a long instance of the interest named in Definition IX, operating through the apparatus named in Definition VI, and, in Guatemala, beside the constraint named in Definition VII. A reader who wants a demonstration that influence creates no desire will not find it in a press tour. He will find it in Part II. A reader who wants a measurement of what advertisements do will not find it in an obituary's sales figures. He will find it in Part IV. A reader who wants to know what the documents support about Edward Bernays will find it here, in the annotations, and will find, in the twelve gaps, the honour of a record that has not been made to say more than it can.

Chapter 30

The Catalogue of Nonsense

The preceding Parts established what influence is and what a technique cannot be. This Part is the empirical warrant. It tests the prediction that follows from Proposition I: if influence is mediated by conditions rather than carried by techniques, then measured effects of persuasion should be small, heterogeneous, and highly sensitive to context, and celebrated findings should fail to replicate once conditions are held constant. Each claim below is given the evidence that can be stated from a compiled audit of the primary literature, then one of five verdicts. This is not a moral trial. The question is not whether the practitioners were wicked. The question is whether the power attributed to them was theirs.

The method of this Part is simple and it is applied uniformly. A claim is stated as it is sold. The original demonstration is named. The replication record, the registered reports, and the meta-analyses that bear on it are reported. A verdict is then issued from a closed set of five: SUPPORTED, WEAK BUT REAL, CONTESTED, REFUTED, HOAX. Empirical claims, which are claims about measured effects, are marked [E]. Documentary claims, which rest on archives, dated publications, and primary testimony, are tried in their own court and are not answered with a statistic. Deductive claims have already been proved in Parts I and II and are not re-proved here. They are used only to say what the measurements mean.

The source of the numbers is a single compiled audit of the influence literature, prepared for this treatise in September 2026. Every citation in that audit was retrieved from primary source metadata or from the primary abstract. Effect sizes are quoted as the named source reported them. Where a number could not be verified from the primary text, the audit says UNSURE rather than supplying a plausible figure, and this Part repeats the uncertainty. All one hundred and thirty digital object identifiers in the audit resolve against the Crossref catalogue to the correct title and year, so the bibliographic layer is machine-verified. Four load-bearing numbers were then re-checked independently against publisher metadata and matched the audit exactly: Tannenbaum and colleagues (2015), d = 0.29 on a composite index, 127 papers, 248 independent samples, N = 27,372; Wagenmakers and colleagues (2016), a rating difference of 0.03 with a 95 percent confidence interval from -0.11 to 0.16; Vohs and colleagues (2021), d = 0.06 with N = 3,531; and Coppock, Hill and Vavreck (2020), 49 advertisements in 59 experiments on 34,000 people. Four further figures were read from the primary PDFs by the compiler of the audit and have not been independently re-verified for this treatise: Hagger and colleagues (2016), d = 0.04; the Open Science Collaboration (2015), a 36 percent replication rate and a mean replication effect of r = 0.197; Sethuraman, Tellis and Briesch (2011), short-term and long-term advertising elasticities of 0.12 and 0.24; and DellaVigna and Linos (2022), 8.7 percentage points against 1.4. Those four are used below with that caution attached. No other numbers are used. A 2020 Nature Communications study of Facebook political advertising, which has circulated in summaries of this literature, does not exist, and it is not cited.

The commercial explanation of the catalogue is postponed until the measurements have been stated, because the explanation is commentary and the measurements are the warrant. One remark is needed at the outset so that the reader does not mistake the arrangement for a polemic. The seller of technique is selling a theory of his own importance. That theory predicts large, portable, easily taught effects. Proposition I predicts the opposite. The catalogue is the test.

Chapter 31

Social and behavioural priming, the Florida effect

As sold, words, images, or stereotypes presented outside awareness activate mental constructs that then drive complex social behaviour. A person primed with words related to the elderly, including the word "Florida", walks more slowly down a hallway afterwards. The demonstration is Bargh, Chen and Burrows (1996), three experiments on politeness and rudeness, on hostility after African-American-face priming, and on walking speed after elderly-stereotype priming. The third is the Florida effect. It entered textbooks, training decks, and the popular literature of automatic influence as proof that a concealed cue can write behaviour.

The walking-speed effect did not survive contact with the original materials under conditions that withheld the hypothesis from the experimenter. Doyen, Klein, Pichon and Cleeremans (2012) failed to replicate it, and found that the effect appeared only when experimenters knew the participant's condition, which is experimenter expectancy rather than automatic behaviour. [E] Harris, Coburn, Rohrer and Pashler (2013) reported two failures to replicate high-performance-goal priming. [E] The Open Science Collaboration (2015) replicated one hundred studies from the psychological literature at large. The compiler of the audit read from that paper a 36 percent replication rate and a mean replication effect of r = 0.197 (standard deviation 0.257), about half the mean original effect size, with 47 percent of original effect sizes falling inside the replication confidence interval and 39 percent subjectively rated as having replicated; 97 percent of the originals had been significant. Those four figures (36 percent, r = 0.197, and the two companion percentages just named) come from the audit's reading of the primary PDF and have not been independently re-verified for this treatise. They describe the climate in which the Bargh claim was re-examined, not a direct re-run of the hallway.

The strongest counterweight is Weingarten and colleagues (2016), a meta-analysis of incidentally presented words. [E] It found a genuine but small behavioural priming effect, d = 0.332 by a fixed-effect model and d = 0.352 by a random-effects model, reported as robust to publication bias. That residue is not the Bargh claim. It is a small effect of incidentally presented words on action, of the sort cognitive psychology has long allowed for masked priming and related phenomena. It does not restore the picture of a concealed stereotype writing a gait, a mood, or a social performance. The large, single-study social priming effects of the Bargh type do not hold. A small, largely automaticity-unrelated priming effect survives meta-analysis. The trade sold the former and, when pressed, points to the latter as if they were the same object.

Verdict. REFUTED. The residual priming in Weingarten, d of about 0.33, is not the Bargh claim.

Chapter 32

Money priming

As sold, exposure to money-related cues makes people more self-sufficient, less helpful, and more accepting of market ideology. The original report is Vohs, Mead and Goode (2006) in Science. An erratum was later issued. The specific effect tested in the Many Labs project was currency priming, from Caruso, Vohs, Baxter and Waytz (2013), on endorsement of free-market systems and social inequality.

[E] Klein and colleagues (2014) tested thirteen effects across thirty-six independent samples totalling 6,344 participants. Ten replicated. Two failed: flag priming influencing conservatism, and currency priming influencing system justification. For currency priming the original effect size was 0.80 (95 percent confidence interval 0.05 to 1.54). The unweighted replication effect was 0.00 (confidence interval -0.06 to 0.09). The aggregate test was t = 0.79, p = .83. Vohs has continued to defend money priming in a 2015 review; the audit saw that manuscript and did not verify its final volume and pages, so the defence is noted and the citation is marked UNSURE. The pre-registered multi-sample evidence is null.

The original effect was large enough to be a product. An effect of 0.00, obtained across thirty-six samples, is the disappearance of the product. The commercial literature did not wait for the second number. It had already converted the first into a story about what money does to the soul, a story that can be told in any room, because every room contains money or the image of it.

Verdict. REFUTED. Many Labs: effect 0.00.

Chapter 33

Ego depletion

As sold, willpower is a finite resource. Exerting self-control depletes it, so a second demanding task is performed worse. The demonstration is Baumeister, Bratslavsky, Muraven and Tice (1998). An earlier meta-analysis by Hagger and colleagues (2010) reported a medium effect, d = 0.62. [E] That figure travelled. It licensed a popular science of decision fatigue, glucose, and the rationing of self-control, and it licensed training programmes that promised to restore a reservoir the first paper had described.

[E] Hagger and colleagues (2016), twenty-three laboratories, total N = 2,141, standardised protocol, pre-registered, found a small effect whose confidence interval included zero: d = 0.04 (95 percent confidence interval -0.07 to 0.15). That d = 0.04 is one of the four figures read from the primary PDF by the compiler of the audit and not independently re-verified for this treatise. [E] Vohs and colleagues (2021), independently re-checked, ran a paradigmatic replication across thirty-six laboratories and N = 3,531. The confirmatory test was not significant, d = 0.06 (95 percent confidence interval -0.02 to 0.14). A Bayesian meta-analysis found the data about four times more likely under the null than under an informed prior of d = 0.30, the size implied by the two large prior multi-laboratory studies. An exploratory analysis without the pre-registered exclusions found a small significant effect, d = 0.08. Friese, Loschelder, Gieseler, Frankenbach and Inzlicht (2019) set out the arguments on both sides. The discipline has not converged on whether a nonzero residual exists.

The medium-to-large effect is gone. Point estimates across the two large registered multi-laboratory studies lie in the range d = 0.04 to 0.08, which is trivial for any practical purpose the trade would recognise as a purpose. A theory of a limited inner fuel, sold as the hidden governor of character, survives as a dispute about a number that cannot be distinguished from zero in the confirmatory tests and that becomes barely visible only when the pre-registered exclusions are dropped. Proposition I predicted that a celebrated finding would fail once conditions were held constant. The registered protocols held them constant. The finding failed in the form in which it had been sold.

Verdict. CONTESTED. Point estimates d = 0.04 to 0.08.

Chapter 34

Power posing

As sold, adopting an expansive posture for a couple of minutes raises testosterone, lowers cortisol, increases risk tolerance, and produces a subjective sense of power. The demonstration is Carney, Cuddy and Yap (2010). It became a public product of unusual reach: a lecture, a posture, a promised change in the body's chemistry, and therefore a promised change in the world's reception of the person who held the posture.

[E] Ranehill and colleagues (2015) ran a sample roughly five times larger (N = 200) and found no effect on hormones or risk tolerance. A small effect on self-reported felt power was present. Simmons and Simonsohn (2017) applied p-curve analysis and concluded that the literature on the embodied effects lacked evidential value. Cuddy, Schultz and Fosse (2018) replied with a p-curve over a broader postural-feedback literature and claimed clear evidential value, though their set is wider than the original claim and includes self-reported outcomes. Carney publicly withdrew her support for the effect in 2016.

Two objects have been confused and must be separated. The hormonal and risk-tolerance claims made the finding a technique. They are refuted. A small effect on the subjective feeling of power remains the live question, and it is not settled. Felt power is an affect already available to introspection. If the posture changes a report of that affect and does not change hormones or risk, what has been shown is closer to a mood induction than to a neuroendocrine intervention. The trade does not sell mood induction. It sells a chemistry.

Verdict. CONTESTED. Hormones and risk are refuted. Felt power is unsettled.

Chapter 35

Facial feedback, the pen study

As sold, holding a pen in the teeth, which forces a smile, makes cartoons funnier; holding it in the lips, which forces a pout, makes them less funny. Facial expression drives emotional experience. The demonstration is Strack, Martin and Stepper (1988). [E] The original reported a funniness-rating difference of 0.82 units on a ten-point scale.

[E] Wagenmakers and colleagues (2016), independently re-checked, collected seventeen independent direct replications under one vetted protocol and obtained a rating difference of 0.03 units (95 percent confidence interval -0.11 to 0.16), which is no detectable effect of the pen manipulation. [E] Coles, Larsen and Lench (2019), a meta-analysis of 138 studies and 286 effect sizes, found the overall effect of facial feedback significant but small, with moderation by outcome type, the presence of emotional stimuli, and stimulus type. Three publication-bias tests found no bias for emotional experience and did find bias for affective judgments. [E] The Many Smiles Collaboration (Coles and colleagues, 2022), a pre-registered multi-centre study of N = 3,878 participants across nineteen countries, found that facial mimicry and voluntary facial action could both amplify and initiate feelings of happiness, but that evidence was less conclusive for the unobtrusive pen-in-mouth manipulation that started the literature.

The pen study, which is the form in which the claim was taught, does not replicate. A broader and smaller facial-feedback effect remains in the meta-analysis and in the multi-laboratory work on mimicry and voluntary action. The honest statement is therefore not that the face does nothing to feeling, and not that a concealed bit of stationery rewrites amusement. The face can participate in an affect already under way. The participation is small, it is sensitive to how the face is moved and to what is present to be felt, and it is weakest in the unobtrusive trick that made the finding famous. That is a WEAK BUT REAL residue, and it is consistent with Proposition II: an influence that appears to create a feeling is operating on a feeling the situation has already made available.

Verdict. WEAK BUT REAL. Wagenmakers: 0.03. The broader literature is small.

Chapter 36

Subliminal advertising, Vicary 1957

As sold, cinemas flashed "Eat Popcorn" and "Drink Coca-Cola" below the threshold of awareness, and sales rose. Subliminal messages can make consumers buy things they did not want. This is the founding myth of the hidden-persuader literature, and it is a documentary claim, not an empirical one. It is answered with the documents.

James Vicary's September 1957 press conference claimed that 45,699 people had been exposed to projections of one three-thousandth of a second, producing a 57.5 percent rise in popcorn sales and an 18.1 percent rise in Coca-Cola sales. In a 1962 television interview with Fred Danzig for Advertising Age, Vicary admitted that the study was "a gimmick" and that the amount of data was "too small to be meaningful." Stuart Rogers (1992) interviewed staff at the theatre and reported the manager saying no such test was ever run. The Federal Communications Commission issued a policy statement in 1974 declaring subliminal advertising techniques contrary to the public interest. Moore (1982) and Pratkanis (1992) reviewed the academic case and concluded that the evidence for subliminal advertising did not support the popular claims. The Pratkanis citation could not be verified from a primary source at the level of volume and pages; the audit marks it UNSURE on those details and not on the existence of the review.

The modern scientific claim is a different object and is tried in the empirical court. Subliminal perception as a phenomenon is real and is standard in cognitive psychology. Masked priming is not in dispute. The narrow advertising claim is much weaker. [E] Karremans, Stroebe and Claus (2006) found that subliminal brand priming affected choice only among participants who were already thirsty. [E] Verwijmeren, Karremans, Stroebe and Wigboldus (2011) found effects only on habitual choices. Subliminal cues appear able to bias actions a person is already prepared to take, not to create new wants.

The founding claim is a HOAX on the author's own admission. The modern advertising claim is REFUTED as a claim about persuasion and real but narrow as a claim about perception. The narrow remainder is a confirmation of Proposition II, and it is as clean a confirmation as the empirical literature supplies. Influence operates on an affect already present. The thirsty person can be biased toward a brand. The person who is not thirsty cannot be made thirsty by a frame of film he cannot see. The trade sold the manufacture of wanting. The laboratory, on the only occasions when it found anything, found selection among wants that were already in force.

Verdict. HOAX (the 1957 claim). Modern subliminal advertising: real only for the already thirsty and the habitual, which confirms Proposition II.

Chapter 37

Neuro-linguistic programming and eye-accessing cues

As sold, people have a dominant representational system, visual, auditory, or kinaesthetic. Eye movements reveal which system is in use. Matching a person's representational language and behaviour builds rapport and allows persuasion. The package has been a commercial curriculum for decades. It has certificates, levels, and a vocabulary that sounds like a science.

Witkowski (2010) reviewed the research database and concluded that the core claims, including eye-accessing cues and representational systems, are not supported. [E] Sturt and colleagues (2012) screened 1,451 papers, found only ten that met inclusion criteria, judged them at high risk of bias, and concluded that there is little evidence that NLP interventions improve health-related outcomes. Eye-accessing cues specifically have not been validated as indicators of representational systems.

Rapport-relevant behaviour has an independent literature. Behavioural mimicry, postural accommodation, and warmth signalling can affect liking and compliance in some settings. That evidence is about rapport and non-verbal synchrony as general social processes. It is not about NLP. It does not validate the package, which is the unit that is actually sold. The trade's method, here as elsewhere, is to attach a proprietary name to a commonplace, to add a diagnostic that does not diagnose, and to sell the resulting bundle as a technology. When the diagnostic fails, the commonplace is offered as the residue, as if the buyer had purchased the commonplace.

Verdict. REFUTED.

Chapter 38

The Mehrabian 7-38-55 statistic

As sold, in all communication 7 percent of meaning comes from words, 38 percent from tone of voice, and 55 percent from body language. What one says does not matter. How one says it does. The statistic is printed on slides, repeated in training rooms, and used to licence a contempt for argument.

What was measured is not what is sold. Mehrabian and Wiener (1967) and Mehrabian and Ferris (1967) gave participants a narrow task. Listeners heard single spoken words, for example "dear", "honey", "maybe", delivered in tones communicating positive, neutral, or negative attitudes, and saw faces expressing positive, neutral, or negative attitudes. The channels were deliberately set in conflict. Listeners judged the speaker's attitude from the channels, and the relative weights were computed from those judgments. The sample was small and drawn from a university population. The 7 / 38 / 55 split describes how much weight listeners gave each channel when the channels disagreed about a single word carrying an attitude about the speaker's feelings.

The generalisation is invalid for reasons internal to the study. The result was about the communication of feeling and attitude, not about information transfer of any kind, and it applied specifically to inconsistent, single-word messages. Mehrabian himself stated that the formula applies only when the communicator is expressing feelings or attitudes and the channels conflict, and that it must not be generalised to other communication. Applying it to persuasion, argument, or instructions is not supported by the study. The statistic, as sold, is a false description of a real and narrow experiment. The false description is the merchandise. The narrow experiment would not have sold a course.

Verdict. REFUTED. A narrow, inconsistent, single-word attitude task was rewritten as a law of all speech.

Chapter 39

Dark psychology, the forty-eight laws, and the art of reading anyone

A neighbouring shelf holds a literature that does not present itself as a laboratory science and should not be tried as one. Books on dark psychology, on forty-eight laws of power, and on how to read anyone offer catalogues of stratagems, historical anecdotes, and promised diagnostics of motive from face, posture, or speech. The audit that supplies the numbers in this Part did not find a replicable scientific object here. No registered multi-laboratory test of a "dark psychology" construct appears in that file. No meta-analysis of the forty-eight laws appears in it. No validated instrument for reading anyone, in the sense the covers promise, appears in it. This Part therefore invents no studies about them, assigns them no effect size, and does not pretend that the absence of a number is a number.

The correct verdict is not REFUTED, because refutation requires a claim stated with enough precision to be false. It is not HOAX, because hoax requires a fabrication of evidence rather than a fabrication of genre. It is unmeasured. The literature is commercial. It has the form of a technique book and the epistemology of a jestbook: a story is told, a maxim is extracted, the maxim is offered as a law. Proposition I already says why such a literature cannot converge. A technique, considered apart from conditions, is not a cause. A catalogue of techniques, considered apart from conditions, is a catalogue of scenes. Scenes can be multiplied without limit. That is why the shelf grows and why nothing on it is retired by a later volume.

What can be said without inventing evidence is only this. In so far as these books repeat claims already tried above, priming, depletion, posing, subliminal influence, NLP, the Mehrabian split, they inherit the verdicts already issued. In so far as they offer maxims unattached to a protocol, they are not scientific claims. The seller is still selling a theory of his own importance. He is selling it without the inconvenience of a method section.

Verdict. Unmeasured. Not a scientific claim. The audit found no replicable scientific object here.

Chapter 40

The Stanford Prison Experiment

As sold, situational roles, not disposition, produce cruelty. Ordinary young men, assigned at random to the parts of guard and prisoner, become their parts. The demonstration is Haney, Banks and Zimbardo (1973), "Interpersonal dynamics in a simulated prison." The citation is the standard one. The journal is not indexed in Crossref, so the audit could not machine-verify the pages. The study is the textbook case of the situation's power, and it has been used for fifty years as if it were an experiment whose result was cruelty.

This claim is documentary. It is not answered with a replication rate. Blum (2018) reported archive documents and participant interviews showing that guards were coached, that the celebrated prisoner breakdown was performed (Douglas Korpi has said he was acting), and that Zimbardo's staff intervened in ways the published account did not describe. A name sometimes attached to the coaching allegations could not be verified as a publication and is not used here. Le Texier (2019) worked directly from the Stanford archives and from interviews with fifteen participants. He documents a biased and incomplete data collection, precise instructions to the guards about how to treat the prisoners, that the guards were not told they were subjects, that participants were almost never fully immersed in the situation, and that the design drew on a prison simulation devised and run by students in one of Zimbardo's classes three months earlier.

A separate experimental critique, Reicher and Haslam's (2006) BBC Prison Study, ran a comparable setting and produced the opposite result: participants rejected the imposed roles and formed a collective against the guards. [E] That opposite result is not a statistical refutation of a historical claim. It is a demonstration that the same theatrical materials, under different instructions and a different investigator, do not yield the textbook cruelty. The archival work is what refutes the original as a report of what happened. The BBC study is what shows that the situation, once deprived of the coaching, does not do what the legend says situations do.

The legend required a clean demonstration that a role writes a character. What the archives show is an investigator writing a role, then writing the report. Recorded here because the industry cites it as evidence that such conditions overpower persons. The conditions that overpowered the prisoners included the investigator.

Verdict. REFUTED. Le Texier, Blum; the BBC Prison Study produced the opposite result.

Chapter 41

Milgram's obedience studies

The original is Milgram (1963), with the 1974 book. [E] In the baseline condition, 65 percent of participants continued to the maximum 450 volts. That is a large effect, and this Part does not pretend otherwise. The question is not whether people shocked a stranger when a man in a grey coat told them to continue. Many of them did. The question is what that fact is an instance of.

[E] Burger (2009) ran seventy adults to 150 volts and then stopped. Because 79 percent of Milgram's participants who passed 150 volts continued to the end, obedience rates could be estimated. Burger found that 70 percent of participants went past the 150-volt point, only slightly below Milgram's comparable rate of about 82.5 percent at that point. A defiant confederate did not reduce obedience. Men and women did not differ. Commentaries by Miller (2009) and Elms (2009) noted that the ethical modifications removed much of what made the original psychologically interesting. The partial modern replication supports the existence of the effect under the truncated protocol.

The interpretation is another matter, and it is tried in the documentary and rhetorical court that the transcripts make available. Packer (2008), in a meta-analytic review of systematic disobedience, found average obedience to 450 volts across conditions of about 43 percent, with variation largely explained by the experimenter's proximity rather than by the victim's. [E] Gibson (2013, 2014, and the 2019 book Arguing, Obeying and Defying) argues from the transcripts that participants were not blindly obeying: many were sceptical, resisted, argued, and negotiated, and the label "obedience" was the experimenter's, not the participants'. A citation of Gibson (2019) as a Journal of Social Issues article is a conflation; the 2019 work is the book. Perry, Brannigan, Wanner and Stam (2020) use an unpublished Milgram test to show that participants who doubted the shocks were more likely to disobey.

The best current estimate in the audit is that, in the baseline paradigm, roughly 60 to 70 percent of participants will continue to the maximum when a legitimate-seeming authority insists, so the effect is real and large. The classic interpretation, blind destructive obedience, is contested by archival work showing that participants acted as active, mostly reluctant, and often sceptical participants in a social situation. Authority, in the sense the trade requires, is a portable lever. What the record shows is a situation of unusual force: a laboratory, a coat, a script, a proximity, and a doubt that, when it is present, predicts refusal. That is consistent with Proposition I. The technique is not the coat. The coat is a condition, and the condition includes the participant's belief that the shocks are real.

Verdict. CONTESTED. The effect is real and large. The blind-obedience interpretation is disputed.

Chapter 42

Asch's conformity experiments

The originals are Asch (1951) and Asch (1956). [E] On the critical trials, roughly one third of responses conformed to the obviously wrong unanimous majority. About 75 percent of participants conformed at least once, which means about a quarter never conformed on any critical trial, and only a small minority conformed on nearly every trial. The audit treats those directional figures as standard and well documented, and notes that the compiler could not re-read the 1956 monograph to quote per-condition percentages exactly. Per-condition figures should be checked against the monograph before being printed as more precise than they are here.

The structure of the effect is the finding. Asch varied the number of confederates. Conformity rises steeply as the majority grows from one to about three and then levels off. Adding more confederates beyond that does little. Introducing a single dissenting ally collapses conformity dramatically. The unanimous majority, not group size as such, carries the effect. Participants also reported substantial distress and private doubt, and many who did not conform were not indifferent to the group. Bond and Smith (1996) meta-analysed the Asch-type literature and found lower conformity than Asch's original, a decline over time in the United States, and higher conformity in collectivist countries. The audit verified those qualitative findings and did not retrieve a mean percentage from the abstract, so no mean rate is quoted. Hodges and Geyer (2006) argue that most participants were resisting a majority they judged arbitrary or incompetent, rather than surrendering their perception.

This is a SUPPORTED effect with a structure, and the structure is what the trade omits. An ally collapses it. A majority of one or two is weak. A majority beyond three adds little. Private doubt is common. A quarter never yield. The effect is not the dissolution of the person in the group. It is a measurable displacement of judgment under a unanimous and immediate social condition, limited by the presence of a single other mind that does not join the error. Axiom IX already said that we are affected by the affects we conceive in those like ourselves. Asch measured one consequence of that fact, in a line-judgment task, and measured also the condition that undoes it. The industry cites the third of critical responses and does not cite the ally.

Verdict. SUPPORTED, with structure. An ally collapses it. About one third of critical responses conform.

The catalogue is a set of claims that share a commercial shape. Each offered a portable cause: a word, a posture, a pen, a glance, a percentage, a role, a coat, a majority. Each, when the conditions were specified or held constant, either vanished, shrank to a residue that is not the thing sold, or survived only with a structure that makes it unusable as a secret. That is what Proposition I said would happen. The next section examines the one taxonomy in the popular literature that was built from real findings rather than from a single demonstration, and asks whether its six names are six laws.

Chapter 43

Cialdini's Six Principles

Cialdini (1984; revised 2007), Influence: The Psychology of Persuasion, lists reciprocity, commitment and consistency, social proof, authority, liking, and scarcity. The principles are real research findings with different evidence weights. They are not six equal laws. The differences matter, and flattening them into a hexagon is the commercial operation performed on a body of work that is better than the hexagon.

Reciprocity is supported, and it is strongest when the gift carries real cost. [E] Regan (1971) showed that an unsolicited favour increased compliance. [E] Falk (2007), in a large field experiment, found that larger gifts produced larger compliance, and that the effect was strongest when the gift was large relative to the ask. The effect is not symmetric. Trivial gifts produce weak effects. This is the opposite of the training-room version, in which a mint or a compliment is credited with the power of a gift. A mint is cheap. A cheap gift is a weak cause. The principle that survives is not "give something, receive something." It is that a costly, unsolicited transfer can move compliance, and that the motion scales with the cost relative to the ask. Cost is a change of condition. The technique, if the word is still wanted, is the cost.

Commitment and consistency are supported and moderate. Freedman and Fraser (1966) demonstrated the foot-in-the-door. Burger (1999) reviewed it and confirmed the effect with modest effect sizes and clear boundary conditions. The audit does not quote a single summary d for this family, and this Part does not invent one. What can be said is qualitative and is already enough to discipline the trade: a small prior assent can raise the probability of a larger later assent, the rise is not large, and it fails under conditions the reviews name. A person is not trapped by a signature. A person who has already acted in a direction is somewhat more likely to act in that direction again. Custom, in the sense of Definition XI, is the longer name of the same fact. Repetition and a prior act are conditions. They are not spells.

Social proof is supported and domain-limited to ambiguity. [E] Cialdini, Reno and Kallgren (1990) show that descriptive norms guide behaviour when they are made salient. The effect is strongest when the situation is ambiguous and the comparison others are similar. Descriptive norms can also backfire when the message advertises how common the undesirable behaviour is. The trade cites the hotel towel and the crowded restaurant and does not cite the backfire. Social proof is not a general solvent. It is a guide used under uncertainty, by people looking at people like themselves, and it can increase the very conduct it was meant to reduce when it is phrased as a census of vice. Ambiguity is the condition. Remove the ambiguity and the proof loses its office.

Authority is the weakest of the six outside Milgram's paradigm. [E] Hofling, Brotzman, Dalrymple, Graves and Pierce (1966) reported that nurses complied with an unknown physician's overdose order. Rank and Jacobson (1977) failed to replicate that finding with a better design. Authority compliance in the Milgram paradigm is well documented and has been treated above. The broader authority principle, the idea that a title, a uniform, or an expert's manner is a portable cause of compliance in ordinary settings, rests on weaker footing. The trade needs the portable cause. The record supplies a laboratory of unusual force and a hospital study that did not come back.

Liking is supported but small and easily confounded. Similarity, praise, and familiarity increase compliance. The effects are typically small and often entangled with reciprocity and commitment. A person is somewhat more moved by a person he already regards with pleasure. Axiom IX said this before the compliance literature measured it. The measurement did not discover a lever. It rediscovered imitation of affect and the endeavour to do what we conceive others to regard with pleasure. The lever, as sold, is a smile and a shared town. The cause is the affect already binding the two.

Scarcity is supported, partly through reactance. Brehm (1966) and Brock (1968) predicted, and Worchel, Lee and Adewole (1975) showed, that scarce objects are valued more highly. [E] The effect is real and is moderated by whether scarcity signals popularity or a threat to freedom. Scarcity is therefore two different conditions under one name. When it signals that others want the thing, it is social proof in another dress. When it signals that a freedom is about to close, it is reactance, which is an affect already present: the endeavour to persist, directed at an action that is being removed. In neither dress is scarcity a technique for creating desire. It is a change in the availability of an object toward which desire already points, or a threat to an action already in the person's power.

The six principles, taken together, are a SUPPORTED taxonomy whose members are unequally evidenced. Authority and liking are the weak links. Reciprocity is strongest when the gift is costly. Commitment is moderate. Social proof is limited to ambiguity and can reverse. Scarcity is real and is not one thing. The honest use of the list is as names for conditions under which already-present affects are more likely to issue in compliance. The commercial use is as six laws a practitioner applies. A law a practitioner applies is a technique. A condition a situation supplies is a cause. Cialdini's better chapters describe conditions. The hexagon on the slide describes a practitioner.

Verdict. SUPPORTED as an unequal taxonomy. Do not flatten them into six laws.

Chapter 44

What Survives Measurement

The residue is smaller than the literature and larger than nothing. It has a shape, and the shape is the shape Proposition I predicted: small effects, high heterogeneity, sensitivity to context, and a handful of robust displacements that do not create desire.

Mere exposure. Zajonc (1968) is the original. Bornstein (1989) meta-analysed 208 experiments and found a robust positive effect, commonly reported as averaging about r = .26, larger for stimuli presented without awareness and following an inverted-U relationship with exposure frequency. The audit could not read that r = .26 from the primary abstract. It is widely reported in secondary sources and is marked UNSURE as to the exact value. What is not unsure is the direction and the existence of the literature: repeated exposure to a stimulus tends to increase liking for it, with limits, and the effect is one of the older findings in social psychology. Repetition is expensive. It requires access to a channel. Definition XI already named the disposition that repetition produces. Mere exposure is the laboratory name of a fact about custom. It is not a secret, and it does not work once. The trade prefers secrets that work once.

Defaults. Johnson and Goldstein (2003) popularised the organ-donation comparison showing that opt-out countries have far higher consent rates than opt-in countries. Jachimowicz, Duncan, Weber and Johnson (2019) meta-analysed default effects and found a reliable average effect with very high heterogeneity. The audit verified the citation and not the numeric average, so no summary d is quoted for that meta-analysis. Counter-evidence on the donation case specifically: Abadie and Gay (2006) found that presumed consent did raise donation but by less than the raw cross-national gap implies, and Arshad, Anderson and Sharif (2014), a panel study of opt-in and opt-out systems, found no consistent mortality or transplantation advantage for opt-out systems after controlling for other factors. Defaults are potent. Their measured size depends heavily on context and on what is being compared. The raw organ-donation gap, which is the form in which the public knows the finding, overstates what a change of default can be shown to do once other differences among countries are admitted. A default is a change of condition. It is among the strongest changes of condition the applied literature possesses. It is also among the least portable, because what a default does depends on what the alternative requires, on what the person already wants, and on what else differs between the settings being compared. The trade cites the gap. The papers that stay with the gap after controls are fewer.

Framing. Tversky and Kahneman (1981) is the original demonstration of gain-versus-loss framing. [E] Klein and colleagues (2014) replicated it in the Many Labs project: original effect size 1.13 (95 percent confidence interval 0.89 to 1.37), replication 0.62 (confidence interval 0.52 to 0.71), with 86 percent of samples in the same direction. The effect replicated and it shrank. Walasek, Mullett and Stewart (2024) found that loss aversion in risky choice is considerably smaller than the standard coefficient usually assumed. The audit verified that article and did not read the abstract, so the claim of a smaller coefficient is drawn from the paper's title and topic rather than from a quoted figure. The magnitude is UNSURE. The direction of the correction is not: the effect is real and has been routinely overstated. Framing is a description of an already-present prospect under one of two aspects. It does not install a prospect. It names the one that is already there in the language of gain or in the language of loss, and the naming moves the rate at which the prospect is accepted. The replication preserved the direction and halved the size. That is the pattern this Part was told to expect.

Anchoring. Tversky and Kahneman (1974) is the original. [E] Klein and colleagues (2014) tested four anchoring effects. All four replicated strongly and in the same direction across all thirty-six samples, with aggregate test statistics such as t = 90.49 and t = 83.66, all p < .001. Anchoring is among the most robust effects in the entire audit. It is also among the most frequently misunderstood. An anchor shifts a number. It does not create a desire. A person who is already going to name a price, a year, or a quantity can be moved toward a number that has been placed in the environment. A person who has no business with that quantity is not given a business by the number. The trade treats anchoring as a method for producing want. The experiments treat it as a method for displacing an estimate. Proposition II is not threatened by a robust anchor. It is illustrated by one. The most reliable technique in the file is not a technique for manufacturing appetite. It is a technique for sliding a figure along a line the person was already on.

Fear appeals. Tannenbaum and colleagues (2015), independently re-checked: 127 articles, 248 independent samples, N = 27,372. [E] The average effect on the composite attitude, intention, and behaviour index was d = 0.29. Effectiveness increased when the message included efficacy statements, depicted high susceptibility and severity, and recommended one-time behaviours. The authors found no circumstances in which fear appeals backfired. This is a real effect of moderate smallness. It is not the omnipotence of fright, and it is not the futility of fright. Fear is an affect. A message that presents an object as dangerous, and that also presents an action as available and useful, can move the composite of attitude, intention, and behaviour by about three tenths of a standard deviation in the literatures assembled under that review. Efficacy is part of the cause. A fear without an action is a different object. Axiom VII said that an emotion is displaced only by a stronger contrary emotion. A fear appeal that works is not information defeating fear. It is one affect, fear plus a usable hope, organised as a message, competing with the affects that were already governing the person. The d = 0.29 is the measured size of that competition in the papers that met the review's rules.

Moral-emotional language. [E] Brady, Wills, Jost, Tucker and Van Bavel (2017), across 563,312 social-media messages on three polarising issues, found that each additional moral-emotional word was associated with a 20 percent increase in diffusion, stronger within ideological networks than across them. This is a large-sample observational association, not an experimental effect. It says that moral-emotional language travels farther in the networks already organised to receive it. It does not say that a moral-emotional word creates a faction, a vote, or a desire. Diffusion is not persuasion. A message that moves farther inside a camp is a fact about the camp and about the words that the camp already treats as tokens of membership. The 20 percent is not a small number for a diffusion study. It is not a number about changing anyone's mind.

Nudges. The applied literature's favourite remainder must be stated twice, because the unadjusted figure and the corrected figure have been sold as if they were one finding. [E] Mertens, Herberz, Hahnel and Brosch (2022), over 200 studies and 450 effect sizes (N = 2,149,683), reported an overall d = 0.45 (95 percent confidence interval 0.39 to 0.52), with moderate publication bias acknowledged. [E] Maier, Bartos, Stanley, Shanks, Harris and Wagenmakers (2022), reanalysing the same data with robust Bayesian publication-bias correction, found that the unadjusted combined estimate of d = 0.43 fell to d = 0.04 (95 percent credible interval 0.00 to 0.14), with a Bayes factor of 41.23 in favour of the null. The two d's, 0.45 and 0.43, are close enough that the correction is not a dispute about which unadjusted paper to believe. It is a dispute about what the file of papers is worth after the file's own bias is modelled.

[E] DellaVigna and Linos (2022) compared 26 academic nudge papers with 126 trials run by two government nudge units. Academic studies averaged an 8.7 percentage-point take-up increase (standard error 2.5). Nudge-unit trials averaged 1.4 points (standard error 0.3). Those two figures, 8.7 against 1.4, were read from the primary PDF by the compiler of the audit and have not been independently re-verified for this treatise. They are used here with that clause attached. The gap is the gap between a literature and a practice. The literature selected, published, and taught the 8.7. The units that run the trials as a matter of government routine obtained 1.4. Benartzi and colleagues (2017) argued that nudges are often more cost-effective than traditional tools. That cost-effectiveness argument can survive a small effect, because a cheap small effect may still be cheaper than a dear large one. The average effect size behind the 2017 framing is far smaller than that framing implied.

Nudges, then, are the qualified case among the survivors. Something in the file is real. After publication-bias correction the combined estimate is d = 0.04, the same order as the registered ego-depletion residuals, and the government units' average is 1.4 percentage points. A default, which is a nudge of a particular material kind, can be potent and is highly heterogeneous. A catalogue of nudges, treated as a portable toolkit, shrinks toward nothing once the file is corrected and once the trials leave the academy. This is Proposition I in administrative dress. The condition (a default, a form, a pre-ticked box) can be a real cause. The toolkit, considered as a set of techniques a unit applies, does not carry the effect the academic abstracts advertised.

What survives measurement is therefore not a secret and not a void. Mere exposure is real and is repetition. Defaults are real, potent, and overstated in their most famous comparison. Framing replicates smaller. Anchoring is robust and does not create desire. Fear appeals, at d = 0.29, are a moderate small effect when efficacy is part of the message. Moral-emotional language diffuses, observationally, by 20 percent per additional word, inside camps. Nudges collapse from an unadjusted d = 0.43 to d = 0.04 after the correction Maier and colleagues applied, and from 8.7 percentage points in academic papers to 1.4 in nudge-unit trials, on the audit's reading of DellaVigna and Linos. The residue is conditions, repetition, access, and already-present affect. It is not a technique.

Verdict. SUPPORTED, with nudges the qualified case: real but small.

Chapter 45

Advertising and Political Persuasion

The trade's largest market is not a seminar on priming. It is advertising, and, in election years, political advertising. If Proposition I is wrong, this is where the wrongness should be most expensive, because this is where the spend is. The verdict is WEAK BUT REAL. The effects are not zero. They are small, noisy, and, in the political case, small across the cuts that were supposed to hide a large effect.

Advertising elasticity. Sethuraman, Tellis and Briesch (2011) meta-analysed 751 short-term and 402 long-term brand advertising elasticities from 56 studies published between 1960 and 2008. The average short-term advertising elasticity is 0.12, substantially below the prior meta-analytic mean of 0.22, and had declined over time. The mean long-term elasticity is 0.24. An elasticity of 0.12 means a 10 percent increase in advertising buys roughly a 1.2 percent increase in sales. Those elasticities, 0.12 and 0.24, were read from the primary PDF by the compiler of the audit and have not been independently re-verified for this treatise. They are used with that caution. Even on that reading, the short-term return is a weak instrument. An increment of 0.12 percent of sales per percent of advertising is not the power the industry's own presentation of itself requires. It is a measurable association between spend and sales, averaged across half a century of studies, smaller than an earlier average, and falling. Long-term elasticity is larger, as would be expected if repetition and custom, rather than a single insertion, carry what carry there is.

Field-experiment returns. [E] Lewis and Rao (2015), twenty-five large field experiments covering about 2.8 million dollars in digital advertising spend, found that the median confidence interval on return on investment is over 100 percentage points wide. The measurement problem is so severe that most realistic advertising effects cannot be statistically distinguished from zero at normal sample sizes. This is not a finding that advertising does nothing. It is a finding that the thing advertising does, if it does it, is smaller than the noise in which it is measured, for spends of the size those experiments used. An industry that cannot distinguish its effect from zero in its own field experiments is an industry that sells attribution in place of demonstration. The attribution is a story about which click caused which sale. The demonstration would be a confidence interval that closed. The interval does not close.

The intention-behaviour gap. [E] Webb and Sheeran (2006) found that a medium-to-large change in intention (d = 0.66) produced only a small-to-medium change in behaviour (d = 0.36). The correlation between intention and behaviour in this literature is approximately r = .36. Persuading someone to intend is not the same as persuading someone to act. The industry measures what it can measure. Intention is cheap to collect. Behaviour is expensive. A campaign that moves a survey item by a detectable amount has not moved the act, and the act is the thing that was paid for. The r = .36 is the empirical form of a fact Spinoza stated without a correlation: an idea of an action is not the action, and an affect is not displaced by a true proposition about what would be better.

Political advertising. Coppock, Hill and Vavreck (2020), independently re-checked, tested 49 real campaign advertisements in 59 experiments on about 34,000 people during the 2016 United States election. [E] They found small average effects on candidate favourability and vote, and, critically, that these small effects were not hiding substantial heterogeneity. Effects were similarly small across sender (candidate or group), receiver partisanship, content type, and context (battleground or not, primary or general). The usual rescue of a small average is the claim that the average conceals a large effect for the right message, the right voter, or the right state. That rescue is what this design was built to test. The rescue failed. The effects were small in the cuts as well as in the mean.

[E] Coppock, Green and Porter (2022) found that a saturated digital advertising campaign moved Democratic vote share by -0.04 percentage points (standard error 0.85). The point estimate has the wrong sign for a persuasion story and an interval that includes a great deal of nothing. [E] Allcott, Gentzkow and colleagues (2025/2026), in Nature Human Behaviour and as NBER Working Paper 33818, ran a field experiment with more than 60,000 participants that removed all political ads from Facebook and Instagram for the six weeks before the 2020 United States election and found no detectable effects on political knowledge, affective polarisation, perceived legitimacy of the election, or participation. There is no 2020 Nature Communications study of Facebook political advertising to cite in place of this, or in addition to it. The verified equivalents are this field experiment and Coppock, Hill and Vavreck (2020). Removing the channel that the popular account treats as a mind-writing machine did not write a detectable change in the outcomes that account names.

Microtargeting. Tappin, Wittenberg, Hewitt, Berinsky and Rand (2023) is the study usually cited for the sentence "targeting adds nothing." The result is more nuanced than that sentence. [E] In the first study, where all messages aimed to move the same policy attitude, their microtargeting strategy outperformed alternatives by an average of 70 percent or more. Targeting on more than one covariate produced no additional gain, and the advantage was visible for only one of the two policy issues. In the second study, where the strategy was used to determine which attitude to target, the advantage was more limited. The defensible summary is that microtargeting can help within a single well-defined messaging task and has not been shown to deliver the large, general persuasive advantage that its vendors claimed. It is not "adds nothing." It is not the vendor fantasy. It is a bounded increment on a task already specified, vanishing when the specification is itself placed in the algorithm's hands, and appearing on one issue rather than on both. Selection, again: the tool is better when the desire to be moved, and the object toward which it is to be moved, have already been supplied.

Cambridge Analytica. As sold, the firm had built psychographic profiles of tens of millions of voters from Facebook data, could identify persuadable individuals, and could target them with advertising tailored to their personality in ways that shaped the 2016 United States election and the Brexit referendum. It also claimed work on campaigns in Kenya, Nigeria, and elsewhere. The academic foundation for psychological targeting is two PNAS papers, and they do not carry the election claim. [E] Kosinski, Stillwell and Graepel (2013) showed that Facebook Likes predict sensitive attributes with high accuracy: 88 percent for male sexual orientation, 95 percent for ethnicity, and 85 percent for political party. Prediction is not persuasion. The paper showed that attributes are predictable. [E] Matz, Kosinski, Nave and Stillwell (2017) ran three field experiments reaching over 3.5 million people and found that appeals matched to extraversion or openness produced up to 40 percent more clicks and up to 50 percent more purchases than mismatched or unpersonalised appeals. Even this supportive result is about clicks and purchases in commercial advertising, not vote choice, and it drew published critiques in PNAS. Prichard (2021) reviews the psychometric basis and the role academic personality research played in enabling the firm's claims.

The firm filed for bankruptcy and ceased operations in May 2018 after the Facebook data-harvesting scandal. No peer-reviewed evidence was ever published showing that its psychographic targeting changed vote choice or election outcomes. The strongest independent estimates of what microtargeting can plausibly achieve are the small and context-dependent returns reported by Tappin and colleagues (2023) and the near-null effects of political advertising in Coppock, Hill and Vavreck (2020) and in Allcott, Gentzkow and colleagues (2025/2026). The gap between the public claims and the evidence is very large. Cambridge Analytica is REFUTED as a claim about changing elections. What is not refuted, because it was shown, is that likes predict attributes, and that personality-matched commercial appeals can raise clicks and purchases. The trade collapsed those three objects into one story. The story was about votes.

The advertising and political file is the industrial form of the catalogue. Short-term elasticity 0.12, on the audit's reading of Sethuraman. Measurement so noisy, in Lewis and Rao, that most effects cannot be distinguished from zero. Intention to behaviour about r = .36. Political advertisements small in every cut. Saturated digital ads at -0.04 percentage points. Removal of Facebook and Instagram political ads without a detectable effect on the outcomes named. Microtargeting a nuance, not a revolution. Cambridge Analytica a refuted election claim built on a real prediction paper and a real click paper. Weak, but real. Real, but not the power that was invoiced.

Verdict. WEAK BUT REAL. Cambridge Analytica, as an election claim: REFUTED.

Chapter 46

The Commercial Explanation

The measurements are now on the table. The commentary that follows is not a fifteenth verdict. It is the account of why a literature of this shape was to be expected, given Proposition I and Definition IX. By Definition IX, the interest of an agent is the direction of its endeavour as determined by its actual conditions. A merchant who tells you that his goods are potent has told you about his situation. The seller of technique is in a situation that requires the belief that technique works. His income, his status, and the intelligibility of his working day are all functions of that belief. He is not required to lie in the ordinary sense. He is required, by the situation, to treat every large effect as a product and every failure as a misuse of the product. Axiom V adds that he is conscious of his desire to be believed and ignorant of the causes that disposed him to that desire. His testimony about his own power is biased, and is insufficient without the comparison Proposition I requires. Part III recorded the documentary form of this fact in the career of Edward Bernays. This section records the empirical form of the same fact in the industry that inherited Bernays's claim and dropped his limit.

Bernays asked whether it is possible to control and regiment the masses according to our will without their knowing it, and he answered, in the same breath, that recent practice had proved it possible, "at least up to a certain point and within certain limits." The industry inherited the claim and dropped the limit. That drop is the commercial operation. A technique with limits is a description of conditions. A technique without limits is a product. The product can be priced, certified, and revised in new editions. The description of conditions cannot, because the conditions (access, repetition, the already-present affect, the absence of an ally, the cost of the gift, the ambiguity of the situation) are usually not the seller's to dispose of, and when they are his they are expensive in a way that cannot be hidden inside a fee for a method.

Who profits from the belief that a technique works is not a mystery and does not require a conspiracy. The publisher profits, because a book that offers a lever is a different object from a book that offers a limit. The trainer profits, because a curriculum of levers can be taught in a weekend and a curriculum of conditions cannot, conditions being local and levers being general. The agency profits, because a client who believes in portable persuasion will purchase volume, targeting, and creative, which are the agency's inventory, rather than purchasing a statement about the client's existing demand, which the client already half knows and does not wish to have invoiced. The platform profits, because a belief in microtargeting is a belief in the platform's inventory of persons. The defence industry, the books that promise immunity, profits from the same belief with the sign reversed: if the lever is real, vigilance is saleable. Part I named this inversion. The paranoid reading is not the opposite of the naive reading. It is the same reading. An agent is credited with a causal power that belongs to the whole.

Failed attempts are explained as bad execution because the alternative is unsaleable. If a priming study fails, the story is that the words were wrong, the hallway was wrong, the experimenters were unskilled. If a campaign fails, the story is that the creative was weak, the targeting incomplete, the budget insufficient. If a nudge fails, the story is that the default was poorly designed. Each of these stories may, in a given case, be true. The systematic function of the story is to protect the theory of the technique. A theory of conditions would predict failure whenever the conditions are absent, and would treat failure as information. A theory of technique must treat failure as error, because a technique that fails when correctly applied is not a technique. The catalogues proliferate for the same reason. A catalogue cannot be wrong. It can only be incomplete. The next volume adds a forty-ninth law, a seventh principle, a new dark art. Nothing is retired, because retirement would require a criterion, and a criterion would require a mechanism, and a mechanism would reduce the catalogue to the handful of conditions this Part has just named.

The honest residue is not saleable as a secret. Defaults work when they change the cost of the alternative, and everyone who has ever set a form knows this without a Greek letter. Repetition increases liking up to a point, and everyone who has heard a song too many times knows the point. Access to a channel is the largest term in any effect that requires the message to be present, and access is rented at known prices. Already-present desire is the condition without which the rest is idle, and it is the one condition the seller cannot admit without admitting that he does not create the thing he is paid to create. Anchoring shifts a number. Fear, paired with a usable action, moves a composite index by d = 0.29. A costly gift moves compliance more than a cheap one. A unanimous majority moves about a third of critical responses and loses them to a single ally. These are the working parts. They are public. They are old. Several of them were stated as axioms before they were stated as findings. A secret must be scarce. This residue is not scarce. What is scarce is the willingness to sell it as residue.

The seller's theory of his own importance has a further property that explains the particular shape of the failures. It requires effects that travel. A finding that holds only in one laboratory, under one script, with one expectancy leaking from one experimenter, cannot be packaged. The Florida effect was packaged. Money priming was packaged. Power posing was packaged. Ego depletion was packaged as a theory of the workday. Each packaging was a claim that the cause lived in the cue, the posture, or the depleted self, rather than in the unmeasured remainder of the situation. Holding the conditions constant is, on that theory, a mere improvement of method. On Proposition I it is the destruction of the appearance. The registered reports did not persecute these literatures. They did what Corollary II of Proposition I required: they compared the same act under conditions that no longer favoured the original appearance. The appearance went.

Publication is a condition of the same kind. A literature that rewards significance will contain more significance than the world contains. Maier and colleagues' correction of the nudge file is the measured form of that fact: d = 0.43 unadjusted, d = 0.04 after the bias is modelled, a Bayes factor of 41.23 for the null. DellaVigna and Linos, on the audit's reading of their PDF, found the same discrepancy between the academic shop window and the government shop floor. The incentive produces a file that looks like a toolkit. The toolkit, used where the incentive to publish a large d is absent, produces 1.4 percentage points.

Cambridge Analytica is the pattern at the scale of a panic. Prediction of attributes from likes is a real paper. Clicks and purchases under personality-matched appeals are a real paper. Votes are a different object. The firm sold the third object. The bankruptcy is not the refutation. The absence of peer-reviewed evidence that psychographic targeting changed vote choice is the refutation. The panic was itself a sale: the defence literature's product, vigilance, moving through the same channel as the firm's product, mastery.

Bernays's modern dress is not a man. It is a relation between a seller and a buyer who need the same false cause. The seller needs it in order to be necessary. The buyer needs it in order to believe that a purchase can install a power the buyer does not already have. Neither need is a conspiracy. Both are interests in the sense of Definition IX. The measurements in this Part are what those interests look like after they have been converted into claims and the claims into protocols. The conversion is lossy. The loss is the distance between d = 0.80 and d = 0.00, between 0.82 rating points and 0.03, between 8.7 percentage points and 1.4, between an election legend and a click.

Influence selects among desires. It does not create them. It operates by changing conditions or by removing countervailing causes. The industry that denies both sentences has a reason for the denial that is not a finding. The reason is the price of the finding.

Chapter 47

The Ledger Restated

Fifteen claims were tried. The table is permitted in this Part because this Part is an audit. Dark psychology, the forty-eight laws, and the arts of reading anyone were examined for a scientific object and none was found; they are unmeasured commercial literature and are not among the fifteen.

Social and behavioural priming of the Bargh type is REFUTED; a small Weingarten residue, d of about 0.33, is not the claim sold. Money priming is REFUTED at 0.00. Ego depletion is CONTESTED, residuals d = 0.04 to 0.08, the first on the audit's reading of Hagger (2016), against a confirmatory d = 0.06. Power posing is CONTESTED: hormones and risk refuted, felt power unsettled. Facial feedback is WEAK BUT REAL: Wagenmakers 0.03, the broader literature small. Vicary 1957 is a HOAX; modern subliminal influence biases the already thirsty and the habitual, confirming Proposition II. NLP is REFUTED. Mehrabian 7-38-55 is REFUTED. Milgram is CONTESTED: a large effect, about 60 to 70 percent in the baseline paradigm, with the blind-obedience reading disputed. The Stanford Prison Experiment is REFUTED by Le Texier and Blum; the BBC study produced the opposite result. Asch is SUPPORTED, with structure: about one third of critical responses, an ally collapsing the effect.

Cialdini is SUPPORTED as an unequal taxonomy: reciprocity strongest when the gift has cost, commitment moderate, social proof limited to ambiguity, authority weakest outside Milgram, liking small, scarcity real via reactance. The replicating cluster is SUPPORTED, nudges qualified: mere exposure, heterogeneous defaults whose organ-donation gap is overstated raw, framing smaller in replication, anchoring robust and number-shifting, fear appeals d = 0.29, moral-emotional language observationally 20 percent more diffusion, nudges from unadjusted d = 0.43 to d = 0.04 after Maier and colleagues, and 8.7 academic points against 1.4 in nudge-unit trials on the audit's reading of DellaVigna and Linos. Advertising and political persuasion are WEAK BUT REAL: short-term elasticity 0.12 on the audit's reading of Sethuraman; Lewis and Rao intervals too wide to distinguish most effects from zero; intention to behaviour about r = .36; Coppock, Hill and Vavreck small in 49 advertisements, 59 experiments, 34,000 people, and not hiding heterogeneity; saturated digital ads -0.04 percentage points; Facebook and Instagram ad removal without detectable effects; Tappin (2023) nuanced. Cambridge Analytica is REFUTED as an election claim. Kosinski predicts attributes. Matz raises clicks and purchases. Neither is a vote.

The record matches the prediction. Proposition I said that a technique, considered apart from a change of condition or the removal of a countervailing cause, is not a cause, and that celebrated findings should fail once conditions are held constant. They failed in that form, or they shrank to a residue that is a condition under another name. Proposition II said that no influence creates the striving, and that a particular desire may be determined from outside only as a determination of an endeavour already present. The only subliminal effects that appear, appear in the thirsty and the habitual. The most robust laboratory effect in the file, anchoring, slides a number along a line the person was already walking. Fear appeals move a composite when a usable action is supplied to an affect that is already fear's kin. Defaults move those who were not going to complete the alternative. Microtargeting helps when the attitude to be moved has already been specified. Advertising's small elasticity is the measured size of an increment to demand that mostly already existed. Nothing in the file requires a fourth power, the manufacture of wanting. The industry is the sale of that fourth power. The sale does not make the power exist.

Part V will rebuild the account that survives: not a catalogue of techniques, and not a void, but the geometry of the residue, the conditions under which already-present affects are displaced, and the limits of that displacement. The industry audited here will not be consulted as a witness to its own necessity. It has been measured. The measurement is the witness.

Chapter 48

The Residue

The audit has done its work. What remains is not a catalogue of surviving tricks, and not a consolation. It is a short list of causes, already named in the Axioms, now restated as Propositions after the inflated account has been removed. This Part is geometric again. The Propositions are [D]. Measurements appear only in scholia, as [E] illustrations of what the demonstrations already require. Documentary remarks in scholia are [H].

"There is no individual thing in nature, which is more useful to man, than a man who lives in obedience to reason."

Spinoza, Ethics, Part IV, Proposition XXXV, Corollary I

A literature can fail in two ways. It can be empty, in which case nothing is left when the claims are withdrawn. It can be inflated, in which case a small true mechanism is buried under a large false one, and the withdrawal looks like a loss of the subject. The influence literature is of the second kind. The residue is real. It is also thin, and it was already available in the Axioms.

The residue is not a set of techniques that happened to replicate. It is a set of operations on affects that the Ethics had already described: fear and hope directed at an uncertain future, the imitation of the affects of those we take to be like ourselves, honour and shame, the pairing of affections by repetition, and the ordinary privation of knowledge about causes. [E] Part IV found small, heterogeneous, context-bound effects wherever a measurement was honest. That is the shape Part I predicted, and it is what a science of affects, rather than a science of scripts, must look like.

This Part rebuilds the account. It does not add new machinery. It derives, from what has already been granted, the few levers that remain after the audit, the reason no appeal defeats a contrary affect already in place, the reason manipulation is not a breach of the order of causes, the reason targeting is selection, the reason fear-based compliance can raise a contrary, the instability of honour, and the one relation between men that is not manipulation. The last of these is the hinge into Part VI.

The rebuilt account can be stated in a page, and that is a mark in its favour. Fear and hope are one machine, because both are inconstant affects of an uncertain future. We imitate those we take to be like ourselves, and we want their regard. Repetition pairs affections and produces a second nature, which is custom. Ignorance of causes is the ordinary condition, not a special wound inflicted by a genius. From these four, together with the definitions of manipulation and propaganda, everything that is true in the trade follows, and everything that is false in the trade is a claim to have added a fifth.

What does not follow is a new optimism. The residue is easier to use than to escape, because it does not require a laboratory, a script, or a theory. It requires a channel, a list, and time. Those are expensive for a critic and cheap for an owner. The book does not end in Part IV, with the bankruptcy of the inflated claims, because a bankrupt literature still leaves the owner of the channel in place. Parts V and VI exist to say what, exactly, he owns, and what a person standing in the channel can still do.

Spinoza did not write a treatise on advertising. Where the application exceeds his text, the text says so.

Chapter 49

A note on what "real" means here

Real, in the title of this Part, is not a compliment and not a metaphysical claim. It means: named as a cause that the Definitions already contain, and not dependent on a finding that failed to replicate. A fear appeal of modest average size is real in this sense. A walking-speed prime is not. The honour of a group is real. A representational system revealed by the angle of the eyes is not. The distinction is the same distinction Part I drew between a causal category and a moral label, now applied after the measurements have been seen.

The reader who wants the residue to be larger will find the want intelligible. A larger residue would justify a larger literature, a larger profession, and a larger fear. The axioms do not consult that want. They name four operations. The audit did not add a fifth. If a fifth is later measured, it will have to be stated as a change of condition, a removal of a countervailing cause, or a new pairing of affections. It will not be stated as a technique that works without those things. That form of statement has been tried, and it is what failed.

Chapter 50

Proposition XIThe operations of organized influence so far demonstrated are four: fear and hope; imitation, including honour and shame; repetition, which is custom; and the maintenance of inadequacy.

Depends on: Axiom VII, Axiom VIII, Axiom IX, Axiom V, Axiom XII, Definition IV, Definition XI, Definition of the Emotions XII, Definition of the Emotions XIII, Definition of the Emotions XXX, Definition of the Emotions XXXI, Proposition II, Proposition V.

Proof. By Proposition II no influence creates the striving; a particular desire may be determined from outside, but only as a determination of an endeavour already present. By Axiom XII the kinds of affect are as many as the kinds of objects. It would seem to follow that the levers are innumerable. They are not. The objects are innumerable; the operations so far demonstrated on an affect, once it is present, are not. By Axiom VIII, the Definitions of the Emotions XII and XIII, and Proposition V, an image of an uncertain future produces hope or fear, which is one operation: the determination of an already striving thing toward or away from an object not yet present. By Axiom IX we imitate the affects of those we conceive as like ourselves and we endeavour to do what we conceive men to regard with pleasure, which is a second operation: the determination of desire by the imagined regard of others, whose names are honour and shame (Definitions of the Emotions XXX and XXXI). Imitation and honour are not two operations; honour is imitation whose object is imagined praise. By Definition XI a repeated pairing of affections produces a disposition, which is a third operation: custom. By Definition IV and Axiom V men act from inadequate ideas of the causes of their own actions as the ordinary case, which is a fourth operation: the maintenance of that inadequacy, by noise, by repetition, or by simple privation. Axiom XI names a different object, the idea of the emotion itself, and is not this fourth operation. Axiom VII does not prove that no fifth operation exists; it says only that an affect yields to a stronger contrary affect. Love, hatred, devotion, and wonder are not hereby excluded as objects. They are, when they move a multitude, determinations of the same four operations, or they remain to be demonstrated as a fifth. These four are the operations so far demonstrated. Nevertheless.

∎

Corollary. A fifth name in the trade's catalogues, whether framing, anchoring, scarcity, or authority, is one of these four operations under another description, or it is not an operation on affect at all.

Corollary II. The appearance of a large number of techniques is an appearance of names. The number of causes has not increased.

Scholium. Part IV supplies the illustration, not the proof. Anchoring, which replicates, is not the manufacture of a desire. It is the determination of a number already being sought, which is to say the attachment of an existing endeavour to a particular magnitude. Defaults, which are potent and heterogeneous, change the condition under which an already present tendency runs, which is Proposition I restated as furniture. Mere exposure is the production of custom by repetition. Fear appeals of modest average size are Proposition V measured. Moral-emotional language travelling inside a network is Axiom IX measured as diffusion. What failed to replicate were the claims that did not name one of these operations: a script that would create a want, a posture that would create a hormone, a depleted will, a hidden syllable that would create a purchase. The audit did not discover a new mechanism. It discarded the names that had never named one.

Consider how the four appear when they are not named. A retirement advertisement shows an old couple on a coast and a number. The number is an anchor. The coast is an image of a future whose issue is in doubt. The doubt is the hope of arriving and the fear of not arriving, which are one machine. The couple is an invitation to imitate people conceived as like oneself, already arrived. The number will be repeated until it is paired with the word security. Nothing in this is a technique in the sense Proposition I relocated. Everything in it is one of the four operations, stacked. Remove the already present fear of a destitute old age, and the stack does nothing. That is Proposition II. Remove the repetitions, and custom is not produced. That is Definition XI. Remove the channel, and the stack is a conversation, which is a different cause.

A wartime poster, a fundraising emergency, a product launch, a feed that rewards moral-emotional language: the stack is the same, in different proportions. The trade's innovation, where it has one, is in the measurement of who already carries which affect, which is selection, and in the cheapening of repetition, which is a fact about capital and machinery. Neither is a discovery about the mind.

The practical consequence is already visible and is costly for anyone who sells catalogues. A catalogue of fifty tactics cannot be learned, transferred, or tested, because it is not a list of causes. A list of four operations can be. The man who asks, of any appeal, which of the four is at work, has already left the trade's literature and entered this one.

Spinoza did not enumerate levers of propaganda. He enumerated the affects and the laws of their combination. The reduction of the trade's catalogue to those laws is the application, and it is ours.

Chapter 51

Proposition XIINo appeal works against a contrary affect already in place. The counter-mechanism is therefore displacement, not resistance.

Depends on: Axiom VII, Axiom X, Proposition IX, E4P14.

Proof. By Axiom VII an emotion can be controlled or destroyed only by a contrary emotion of greater power. By E4P14 a true knowledge of good and evil cannot check any emotion by virtue of being true, but only in so far as it is considered as an emotion. By Axiom X the knowledge of good and evil is nothing else but the emotions of pleasure or pain, in so far as we are conscious thereof; the language of resistance, of seeing the right and holding to it, is therefore a report of affect in the guise of a faculty. An appeal, whether a demonstration, a warning, a frame, or a script, is an attempt to determine an affect. If a contrary affect is already in place and is the stronger, the appeal, considered as information, does nothing, and considered as an affect it is the weaker of the two. By Proposition IX this is the case in which no technique suffices. The only remaining operation is the supplying of a contrary affect stronger than the one in place, which is displacement. Resistance, understood as the holding of an affect at bay by a faculty that is not itself an affect, is excluded by Axiom VII and Axiom X. Nevertheless.

∎

Corollary. Advice that consists in the instruction to resist, to be vigilant, or to refuse a frame, is either the production of a contrary affect (fear of being taken in, pride in not being taken in, shame at compliance) or it is nothing.

Corollary II. An education that increases causal knowledge without supplying a contrary affect increases the instability of a manipulation (Proposition IV) without guaranteeing its end.

Scholium. The popular defence literature is a literature of resistance. It pictures a man who sees the technique and, seeing it, is no longer moved. The picture is false twice. First, seeing a technique is not the same as being affected by a stronger contrary cause; a man may name the scarcity cue and still buy, and his naming is then a further humiliation rather than a liberation. Second, the seeing, when it does work, works because it is itself an affect: contempt for the seller, fear of looking foolish, the honour of being the one who was not taken in. Those are displacements. They are real. They are not the faculty the literature advertised.

The same error appears on the other side, among those who imagine that a true account of a product, a war, or a candidate will of itself correct a passion. It will not. The true account must become an affect, or it must wait upon one. Public health campaigns that add an efficacy statement to a fear appeal are, whether they know it or not, attempting to supply a second affect (hope of a practicable act) rather than to argue the first one out of existence. Tannenbaum's result that fear appeals work better with efficacy statements is a measurement of this Proposition, not a discovery of a new one.

What follows for conduct is reserved to Part VI. What follows for theory is immediate. Anyone who promises a defence that is not the construction of a contrary affect has promised a violation of Axiom VII, and there are no such violations.

Chapter 52

Proposition XIIIManipulation is not a violation of nature but a case of it. A conspiracy of modes, if it occurs, is still nature operating.

Depends on: Axiom III, Axiom V, Axiom VI, Definition IV, Proposition III.

Proof. By Axiom VI inadequate and confused ideas follow by the same necessity as adequate ones, and falsity is a privation of knowledge rather than a positive entity planted from outside. By Definition IV and Proposition III manipulation is influence in which the affected person's idea of the cause of his own action is inadequate. Such an influence is therefore an instance of the same necessity that produces any other inadequate idea. By Axiom III nothing is contingent. A crowd moved by a false account of its cause is not an exception to the order of nature, a lapse, or an invasion of a kingdom within a kingdom. It is nature operating. By Axiom V the operator is as ignorant of the causes of his own striving as the subject is of his, so the image of an unmoved mover standing outside the order and arranging it is excluded twice: once as a violation of necessity, and once as a violation of the operator's own condition. Organized bodies are already named in Definition VI. Men who meet, agree, and repeat are still modes, still caused, still ignorant of a share of their own striving. What remains is a network of causes, some of them organized, none of them exempt. Nevertheless.

∎

Corollary. The search for a hidden master who explains every mood is the same error as the search for a hidden faculty that vetoes every mood. Both credit an agent with a power that belongs to the whole.

Corollary II. To name a man as a cause of a repetition, a channel, or a fear is legitimate. To name him as the cause of the order of causes is not.

Scholium. This Proposition is the one that will be most resented, because it appears to abolish blame, and because the appetite for a conspirator is itself an affect (a fear that has found an object, and a hope that the object, once named, can be removed). It does not abolish blame. It relocates it. A man who sells a false account of his own power, who repeats a fear forty times a week, who owns a channel and uses it to occupy the space in which a cause might become clear, is a cause, and it is proper to say so. What is improper is the further step, in which that man is made to bear the causal load of the hungers, humiliations, customs, and ignorances he did not produce and does not understand. The further step is comforting. It is also a mistake, and a mistake about causes cannot guide anyone's action.

The documentary Part has already shown the pattern in a single career. Bernays's interest determined his account of his own efficacy (Proposition VIII). The account was inherited as a theory of history: a nephew of Freud, a hidden government, a century of the self. The documents, where they can be checked, show something smaller and more ordinary: a state training-ground, a roster of clients who already owned presses and radio, a partner who wrote the sentences, a memoir that enlarged the rest. The enlargement is not a conspiracy. It is Axiom V applied to a man whose product was the claim that he stood outside Axiom V.

Spinoza's sentence is the whole of the political counsel that follows from this Proposition: men are not a kingdom within a kingdom. The paranoid inversion, which makes them a kingdom under a hidden king, has not escaped the error. It has changed the monarch.

Chapter 53

Proposition XIVThe effect of any given appeal is heterogeneous across persons. Targeting is therefore a matter of selection, not of technique.

Depends on: Axiom XII, Proposition II, Proposition IX, E3P56.

Proof. By Axiom XII and E3P56 there are as many kinds of pleasure, pain, and desire as there are kinds of objects by which we are affected. By Proposition II an appeal determines an endeavour already present; it does not create the striving. An appeal that names a given object will therefore move those in whom that object already determines a desire, and will not move those in whom it does not. The difference is a difference in the persons, not in the appeal. What is called targeting is the selection of those persons in whom the relevant affect is already in force, or the withholding of the appeal from those in whom a contrary affect is in force (Proposition IX). It is not the alteration of the appeal into a universal instrument. A universal instrument is excluded by the heterogeneity of objects. Nevertheless.

∎

Corollary. A large average effect of an appeal, reported without a measure of who was already disposed, is not a measure of the appeal. It is a measure of the sample.

Corollary II. The failure of a celebrated technique to replicate, once the sample is no longer tacitly selected for the relevant affect, is the predicted result of this Proposition and not an anomaly.

Scholium. The trade reports selection as persuasion because persuasion is what it sells. A list purchased because it contains people who already bought, a lookalike audience built from those who already clicked, a fear appeal sent to those who already watch the weather, a patriotic image sent to those who already stand for the anthem: each of these is selection. The script can be held constant and the effect will still appear, which is why the script is credited. Hold the persons constant and vary the script, and the credit moves. Part IV's small, context-bound, highly heterogeneous effects are what this Proposition requires. Coppock, Hill and Vavreck found that the small effects of political advertisements were small across sender, receiver, content, and context, which is to say that the sample of voters, already sorted by a lifetime of prior affections, did not contain a hidden pocket of persons waiting to be made. Where an effect is large, look first for the prior affection that was not measured.

Microtargeting, in the form sold to campaigns, is this Proposition converted into a product. The honest version of the product is: we will try to find the people who already want the thing. The dishonest version is: we will write a sentence that creates the want in whoever receives it. The first version is selection. The second is the portable cause that Proposition I relocated. The gap between them is the industry's margin.

Spinoza did not discuss targeting. He discussed the heterogeneity of the affects. The rest is the application.

Chapter 54

Proposition XVOnce selection and repetition are named, what remains of a mass-media effect is a change of condition or the removal of a countervailing cause.

Depends on: Definition VI, Definition XI, Axiom II, Proposition I, Proposition VII, Proposition XIV.

Proof. By Definition VI propaganda is organized, collective, addressed to a multitude, and sustained by repetition. By Proposition VII access is a necessary part of the definite cause at that scale. By Proposition XIV an appeal's effect is a matter of who already carries the relevant affect. By Definition XI repetition produces custom, which is a disposition of body and mind. By Axiom II the order of ideas is the order of things, so the material facts of who is reached, how often, and through what channel are not contexts of a persuasion. They are causes. What remains, once selection and repetition are named, is the residual that the trade calls persuasion: the contribution of the particular form of words, considered apart from who was selected and how often the words arrived. By Proposition I that residual is either a change of condition or the removal of a countervailing cause. The demonstration stops there. Whether the residual is small when the two named terms are large is an empirical question, marked [E] where it is answered. Nevertheless.

∎

Corollary. A campaign that cannot name its selected population and its repetitions has not specified its causes, and a report of its success is a report of a scene.

Corollary II. The same words, deprived of the channel and the list, are a different cause, and no transfer of effect from the first case to the second is licensed.

Scholium. This is the Proposition that converts the four questions of Part I into a claim about mass media. The first question (what affect was already present) is selection. The third (who controlled the channel and the repetition) is distribution. The second (what condition changed) is whatever remains. The fourth (what the person understands) decides whether the case is manipulation. A newsfeed is a selection engine before it is a persuasion engine. A brand that buys saturation is buying custom. A state that owns the only station is not a better rhetorician than a state that does not; it is a monopolist of repetition.

[E] The measurements illustrate. Advertising elasticities on the order of a tenth (compiler-read, Sethuraman), political advertisements that do not move vote share once they are tested rather than remembered, the collapse of published nudge effects once publication bias is corrected: these are what one finds when selection and repetition are no longer allowed to hide inside the word persuasion. They are not proofs. They are the shape the record takes if the proof is true.

A warning, which is also a limit. This Proposition does not say that words do nothing. Words are modifications of body and mind, and they can be the occasion on which a desire becomes salient, an object is attached, or a countervailing fear is interrupted. It says that those occasions, in the mass, are not where the power is. The power is in who is reached, how often, and what they already were. The trade cannot sell that sentence, because the buyer of the trade already knows who he can reach, and does not wish to be told that this, and not the copy, is the purchase.

Chapter 55

Proposition XVIFear-based compliance may, when the fear becomes a common affect, raise a contrary solidarity, and is in that case self-limiting.

Depends on: Axiom VII, Axiom IX, Proposition V, Definition of the Emotions XIII, E4P63.

Proof. By Proposition V, where propaganda has as its object an unsettled future, the lever that matches its duration is an affect of the imagination directed at that future. Fear is that affect on its painful side (Definition of the Emotions XIII). By E4P63 he who is led by fear, and does good in order to escape evil, is not led by reason. The compliance so produced is real. It is also a pairing: the persons who share the fear become, by Axiom IX, objects of imitation for one another, and the fear is thereby increased as a common affect. A common fear of an external object may produce a common endeavour against that object, which is often the propagandist's aim, and is solidarity with him so far as he is taken as the means of escape. The same common affect may, under further conditions, be attached to hatred of whoever appears to traffic in the fear, and that solidarity is no longer the propagandist's. Axiom VII permits the second attachment; it does not necessitate it. When it occurs, the instrument has raised up a contrary cause. The proposition is therefore a limit that may be reached, not a law that every successful fear campaign must reach. Nevertheless.

∎

Corollary. A fear that cannot be shared, because it isolates, does not produce this solidarity and is limited instead by exhaustion, numbness, and the ordinary decay of an image that is never confirmed.

Corollary II. Hope used without fear is not available as a durable lever, because by the Definitions of the Emotions XII and XIII hope and fear are one machine. What is called hope in a campaign is fear of the loss of the hoped-for thing.

Scholium. The self-limitation is visible wherever a panic has lasted long enough to become a people. A wartime public bound by fear of an enemy will, if the fear is successful, come to love the others who share it and to hate whoever appears to traffic in it. The trafficker then has a new problem, which is not disbelief but a solidarity that can be turned. Occupations discover this when the fear that produced quiet begins to produce a resistance whose members find one another by the mark of the same fear. Platforms discover a smaller version when a threatened group becomes an audience that will punish the platform for the threat. None of this requires a theory of awakening. It requires Axiom VII and the imitation of affects.

Spinoza's political writings treat superstition as a use of hope and fear. He did not write the sentence that fear-based compliance may raise a contrary solidarity. The sentence is the application of E4P63, Axiom VII, and Axiom IX to organized fear, and it is offered as such.

The limit is not a consolation to the person presently afraid. An affect is not removed by the news that it will one day raise up its own contrary. It is removed by a stronger contrary affect now. Part VI will not pretend otherwise.

Chapter 56

Proposition XVIIThe desire for approval attaches to an object already installed and not owned by the operator, and is therefore both ready to hand and unstable.

Depends on: Axiom IX, Definition of the Emotions XXX, Definition of the Emotions XXXI, Proposition X, E4P58S.

Proof. By Axiom IX we endeavour to do what we conceive men to regard with pleasure, and shrink from what we conceive them to shrink from. Honour is pleasure accompanied by the idea of an action of our own which we believe to be praised by others (Definition of the Emotions XXX). Shame is the corresponding pain (Definition of the Emotions XXXI). These affects require no new object to be installed in the affected person. The object is already there: the imagined regard of others. An organized influence that attaches its preferred act to that regard, or its forbidden act to the corresponding blame, therefore operates on an endeavour already present (Proposition II). That is the readiness. By Proposition X a successful manipulation produces obedience without adequate ideas, and is fragile: the same affect can be recaptured by any stronger cause. Honour and shame are such affects, and they have this further fragility, that their object is the opinion of others, which is itself movable by a competing influence. The operator does not own that object. Spinoza notes that honour is not repugnant to reason (E4P58), and, in the scholium, that empty honour, rooted in popular approval, must wander with that approval (E4P58S). The instability is paid by the operator: he has used a ready compliance and has not acquired the stability of the cause. Whether this is cheaper in money than a new fear is an empirical question, marked [E] where it is answered. Nevertheless.

∎

Corollary. An institution that governs primarily by honour and shame must continually produce the spectacle of regard, because the affect decays when the imagined audience is not refreshed.

Corollary II. The person who lives for the regard that the lever exploits is the most available to it and the least able to state the cause of his own action, which is to say the most manipulated in the sense of Definition IV.

Scholium. Honour is ready to hand because it is already installed. Every school, army, platform, office, and church has discovered this, and none of them needed a laboratory. The like-button, the honour roll, the denunciation, the prize, the quiet exile from the group: these are not techniques in the sense Proposition I relocated. They are attachments of an existing endeavour (to be regarded well by those we take as like ourselves) to particular acts. The cost of production is the cost of making the regard visible. [E] In ordinary institutional practice that cost is smaller than the cost of installing a new fear, which requires an object, a future, and enough repetition to keep the image present. The comparison is not a step in the demonstration.

The price is less often counted. A brand that lives by cool is captured by whoever next defines cool. A party that lives by the shame of the unfaithful is captured by whoever next defines faith. A man who has been governed by the honour of his trade will, when the trade's honour is redefined, be governed still, and will call the new obedience loyalty. Proposition X is the general case. Honour is the case that costs the least to begin and the most to escape, because escaping it is itself shameful in the eyes of the audience that installed it.

E4P58 is Spinoza's, and it is not a condemnation of honour. Honour that follows from adequate ideas, from the desire to live in a manner that a free man can understand, is not the lever. The lever is empty honour, the pleasure of being praised by a multitude whose reasons one cannot state. The trade uses the second and advertises it as the first.

Chapter 57

LemmaFor a finite mode, freedom is the increase of the proportion in which the mode is the adequate cause of its own action.

Depends on: Definition III, Definition XII.

Proof. Definition XII is E1D7: that thing is called free which exists solely by the necessity of its own nature and is determined to act by itself alone. For a finite mode this is a limit case, as already stipulated. What follows for a man was left to be derived. By Definition III we act when we are the adequate cause of what takes place in us or through us. The only increase available to a finite mode, short of the limit case, is therefore an increase in the proportion of its actions of which it is the adequate cause. That increase is what this book will mean by human freedom. It is not a second definition of freedom. It is the content Definition XII deferred.

∎

Chapter 58

Proposition XVIIINothing is more useful to man than a man who lives under reason. Influence between such men is persuasion, not manipulation.

Depends on: Definition IV, Definition V, Proposition III, E4P35, E4P35 Corollary I, E4P35 Corollary II, Lemma of this Part.

Proof. By E4P35 men agree in nature only in so far as they live in obedience to reason. By the first corollary, there is no individual thing more useful to man than a man who lives in obedience to reason. By the second, men are most useful to one another when each seeks what is useful to himself under reason. By Definition V persuasion is influence whose effect is an adequate idea of the cause of the affected action. By Definition IV and Proposition III manipulation is influence whose effect is the opposite. Influence between men who understand the causes of what they ask and of what they do is therefore persuasion, and it increases the power of acting of both, which is the sense of usefulness in the corollary. Such influence may be repeated and shared. It is not propaganda in the sense of Definition VI unless it addresses a multitude determined by affects rather than by reason, which is the contrary of the E4P35 case. It is not exemption from being moved. It is the increase of adequate causation in the moving, which is the Lemma of this Part. The opposite of manipulation is not isolation. Isolation would be the refusal of the most useful thing. The opposite is understanding. Nevertheless.

∎

Corollary. A community whose members can state the causes of their common actions is, to that extent, not a multitude in the sense of Definition X, even if it is numerous.

Corollary II. The attempt to escape influence by withdrawing from other men withdraws from the principal aid to understanding, and so tends to increase, rather than diminish, inadequacy.

Scholium. This Proposition is the one the defence literature cannot write, because that literature has defined safety as the absence of being moved. Spinoza defines freedom as the quality of the moving. A teacher who shows a cause, a physician who explains a mechanism, a colleague who names the incentive that is governing a meeting, a friend who can say why a fear has returned: these are influences. They are the useful ones. They do not require a purity of intention in the teacher, the physician, the colleague, or the friend, because Definition V does not consult intentions. They require that the affected person be able to state the cause.

The trade will try to occupy this sentence. It will call its own work education, empowerment, thought leadership, authentic conversation. The test is not the name. The test is Definition III: can the action be understood through the agent's own nature alone? Recitation is evidence in one direction only. A man who can state no cause at all is giving evidence that his idea is inadequate. A man who can state causes has not thereby proved his idea adequate (Definition V, Note). The occupation of the word is Axiom X applied to marketing.

Part VI takes up the conduct that follows. The chain is now complete enough for that: the operations so far demonstrated are few, resistance is a misunderstanding of Axiom VII, there is no unmoved mover who stands outside the order, targeting is selection, the residual of a mass effect is a change of condition, fear may raise a contrary, honour is ready and unstable, and the contrary of manipulation is not solitude. What remains is to say what a man can do who wishes to be, not unmoved, but more nearly the cause of his own motion.

Chapter 59

Lemmas on the Counterfactual

The demonstrations above treat the typical case. The same operations, under different material conditions, produce different magnitudes. The lemmas record that fact so that it is not later used as an objection.

Lemma I. The same form of words, addressed to a person in whom the relevant affect is absent, is not the same influence as those words addressed to a person in whom it is present. By Proposition XIV the first is a failed selection. It is not a weaker instance of the second.

Lemma II. The same form of words, deprived of repetition and of a channel, is not the same influence as those words under saturation. By Proposition XV and Definition XI, custom has not been produced. A laboratory exposure is not a campaign, and a campaign is not a conversation.

Lemma III. The same fear, where a practicable act is attached to it, is not the same affect as that fear without an act. The first is hope-and-fear as one machine. The second is fear tending toward numbness or toward a solidarity that the user does not own (Proposition XVI).

Lemma IV. The same honour, attached to an act whose cause the honoured person can state, is not the lever of Proposition XVII. It is closer to the honour Spinoza excepts from emptiness. The test is not the presence of praise. The test is whether the praise could be removed without the act failing.

These lemmas are the reason the book refuses transfer. A finding about a classroom, a website, a wartime public, or a brand cannot be carried into another of those settings by the name of the tactic. The setting is the larger part of the cause. Anyone who has followed the chain this far already knows why.

Chapter 60

Proposition XIXThe remedy for manipulation is an adequate idea of the cause of one's own action.

What follows is not a set of techniques for remaining unmoved. That condition does not exist, and the promise of it is one of the manipulations this book has described. What follows is derived from the Propositions already demonstrated. Each protocol cites its Proposition. The aim is an increase in the proportion of one's actions of which one is the adequate cause. The aim is not a completion.

"A free man thinks of death least of all things; and his wisdom is a meditation not of death but of life."

Spinoza, Ethics, Part IV, Proposition LXVII

The defence literature ends with a kit. The kit is a list of cues to spot, frames to refuse, and inner speeches to recite. The kit presupposes the faculty Part I denied, and it recruits the fear Part II named. A man who has bought the kit has been sold a new object for an old affect. He is not less moved. He is moved by the trade that claims to unmove him.

This Part does something else. It takes the demonstrations as given and asks what conduct they require. Because an affect is removed only by a stronger affect, the conduct cannot be a set of reminders. It must be the construction of contrary affects whose objects are the true causes. Because manipulation is a relation to one's own knowledge, the conduct cannot be a hunt for operators. It must be an increase in the adequacy of the ideas from which one acts. Because the levers are few, the conduct can be general. Because the effects are heterogeneous, the conduct cannot be a list of situations.

Spinoza did not write these protocols. He wrote E5P3, E5P4, E4P7, E4P14, E4P67, and the definitions of action and freedom. The protocols are the application, and they are offered as such.

Depends on: Definition IV, Proposition IV, Proposition III.

Proof. By Definition IV and Proposition III the manipulation-relation holds in the same degree as the inadequacy of the affected person's idea of the cause of his own action. By Proposition IV every manipulation is unstable in proportion to the causal knowledge available. Supply an adequate idea of that cause, and the relation holds of less of the influence, and, in the limit, of none of it. No other remedy reaches the relation named in Definition IV, because that relation is the privation of such an idea. Separately, by Axiom XI, an emotion which is a passion ceases to be a passion as soon as we form a clear and distinct idea of the emotion itself (E5P3), and there is no modification of the body of which we cannot form some such conception (E5P4). That cessation concerns the affect as a passion. It is not the remedy for the manipulation-relation, whose object is the idea of the cause of the action. The two may accompany one another. The second is not carried by the first. Nevertheless.

∎

Corollary. A remedy that does not increase the affected person's idea of the cause of his own action may change his conduct and still leave him manipulated, if he still cannot state the cause.

Corollary II. The remedy is available in degree. A clearer idea of the cause of the action is a less complete manipulation-relation. Completion is not required, and by E2P23 is not available as a finished state of self-knowledge.

Scholium. Causal knowledge, in this use, is not information about the operator. A man may know the name, salary, and method of the person who produced a campaign and still have no idea why he, in particular, was available to it. The cause that Definition IV requires is the cause of his own action: the affect that was already present, the condition that changed, the repetition that installed the object, the honour or fear that carried the act. The operator is one term, and often not the largest.

Nor is causal knowledge a guarantee. Corollary II of Proposition IV already said this: by E4P14 a true idea does not check an emotion by being true. The manipulation-relation lessens when the idea of the cause is formed; the conduct may still follow the old pairing until a stronger affect displaces it. Knowledge is the remedy for manipulation as a relation. It is not, by itself, the remedy for a habit. The protocols below therefore pair the formation of the idea with the construction of a contrary affect. Either alone is incomplete.

Chapter 61

Proposition XXVigilance is not the remedy named in Proposition XIX. It is displacement by fear.

Depends on: Axiom VII, Axiom VIII, Proposition XII, Proposition XIX, Definition IV.

Proof. By Proposition XIX the remedy for the manipulation-relation is an adequate idea of the cause of one's own action. Vigilance, as sold, is a posture of watchfulness toward techniques, a readiness to refuse frames, a suspicion of being moved. That posture is an affect: a fear of an uncertain future harm (being taken in) directed at an image of hidden operators. By Axiom VIII the image works as a present thing. By Axiom VII this affect can be stronger than a given appeal, and can therefore displace it, which is why vigilance sometimes appears to work. It has not worked as the knowledge Proposition XIX requires. It has worked as a contrary fear. By Proposition XII that is displacement. Displacement can be a remedy for a particular appeal, in the only sense Proposition XII allows. It is not the remedy for the relation named in Definition IV, because the person who has been moved to refuse may still have no adequate idea of the cause of the refusal. The sellers of defence, of conspiracy, and of purity all produce that fear. A change of object that leaves the inadequacy in place is a change of manipulator, not an increase of adequacy. Nevertheless.

∎

Corollary. A life organized around the detection of techniques is a life organized around a fear, and will be available to whoever next names the threat.

Corollary II. The test of a supposed defence is the same as the test of any influence: can the person state the cause of his new refusal, or has he only acquired a new object for an old passion?

Scholium. The paranoid inversion of Part I is the completed form of vigilance. Every mood is assigned an author. Every purchase is a defeat. Every kindness is a tactic. The assignment feels like knowledge because it is specific, and it is not knowledge because it does not name the causes in the affected person. It names an enemy. An enemy is an excellent object for fear and for honour (the honour of being the one who saw). Proposition XVII applies. The cheapest lever has been turned upon the person who thought he had escaped levers.

There is a saner vigilance that is not this Proposition's target: the ordinary attention of a person who asks the four questions of Part I. That attention is a method of forming ideas of causes. It does not require a posture, a kit, or an enemy. When the method hardens into a posture, it has become the affect this Proposition names.

Chapter 62

Proposition XXIThe free man is not the man who is unmoved. He is the man who is the adequate cause of his own motion.

Depends on: Definition III, Lemma of Part V, Proposition XIII, E3P1.

Proof. By Definition III we act when we are the adequate cause of what takes place in us or through us, and we are acted upon when we are only a partial cause. By the Lemma of Part V, human freedom is the increase of that adequacy, not exemption from being caused. By Proposition XIII there is no position outside the order of causes, so the unmoved man is not a possible case. By E3P1 the mind is active so far as it has adequate ideas and passive so far as it has inadequate ones. Freedom is therefore a proportion, reversible, and never complete. Nevertheless.

∎

Corollary. A practice that aims at not being influenced aims at an impossibility and will produce, in its place, the fear of influence.

Corollary II. A practice that aims at knowing the causes of one's present motion aims at the only freedom available to a mode, and will produce, in so far as it succeeds, an increase of power to act.

Scholium. E4P67 states an orientation, not a step in the proof: a free man thinks of death least of all things, and his wisdom is a meditation of life. In the present subject that sentence is useful as a warning against a practice organized around the fear of operators. It is not a demonstration that meditation of death equals fear of propaganda.

The free man of this Proposition still laughs at jokes, still buys bread because he is hungry, still is moved by a death, a slight, a proof, a face. He is not a stone. The difference is whether a greater share of what he does can be understood through his own nature (Definition III). Recitation may be evidence of inadequacy; it is not the measure of adequacy. He will be wrong about some of it, because by E2P19 and E2P23 the mind has no adequate knowledge of the body or of itself except through the ideas of its affections. The Proposition does not require a completed inventory. It requires a direction: away from the fantasy of exemption, toward the increase of adequacy.

This is also why the book has no chapter on becoming unmanipulable. That chapter would be a lie about the Lemma of Part V. What it has is a chapter on becoming, in a larger proportion of cases, the cause from which the action can be understood.

Chapter 63

Proposition XXIIBecause affects are removed only by stronger affects, the practical measure is the construction of a contrary affect at the moment of exposure.

Depends on: Axiom VII, Proposition XII, Proposition XIX, E4P14.

Proof. By Axiom VII and Proposition XII resistance is unavailable and displacement is the only counter-mechanism. By Proposition XIX the remedy for the relation of manipulation is a clear idea of the cause. By E4P14 that idea does not check the emotion by being true. It must itself be, or be joined to, an affect of greater power. The practical measure is therefore not the storage of maxims against a future appeal. Maxims stored in the absence of the appeal are ideas without the contrary affect, and they fail in the presence of the appeal for the reason already demonstrated. The measure is the construction, at the moment of exposure, of a contrary affect whose object is the true cause: not the product, the candidate, or the threat as presented, but the pairing, the repetition, the honour, the already present hunger or fear that made the presentation land. Nevertheless.

∎

Corollary. A contrary affect whose object is still the presented thing (contempt for the product, hatred of the candidate) has not displaced the mechanism. It has changed the sign of the same object.

Corollary II. The construction cannot be guaranteed in advance, because the relative power of two affects is a fact about the moment and not about a resolution taken in another mood.

Scholium. The moment of exposure is ordinary. It is the third repetition of a headline, the silence after a slight, the cart, the feed, the brief panic at a number. In that moment the available contrary affects are few, and they are not noble. They include: the shame of repeating oneself, the weariness of a known pairing, the honour of not performing for an unseen audience, the fear of a later accounting one actually believes, the pleasure of a different object already at hand. The protocol is not to pick the most virtuous of these. It is to use whichever is stronger and is pointed at the cause rather than at the bait.

Spinoza's own remedies in Part V of the Ethics (the ordering of the body's affections, the joining of an affect to a true cause, the love of God as the most constant affect) are the general form of this Proposition. He did not apply them to a newsfeed. The application is ours, and it is narrower: we do not claim the intellectual love of God as a consumer skill. We claim only that a contrary affect, now, pointed at the cause, is the available act.

Chapter 64

Proposition XXIIIGiven the heterogeneity of affects, the adequate defence is general and structural, not situational and tactical.

Depends on: Axiom XII, Proposition XIV, Proposition XV, Proposition XXII, Definition XI.

Proof. By Axiom XII and Proposition XIV no single appeal has a uniform effect, and targeting is selection. A defence composed of tactics matched to tactics (when you see scarcity, do this; when you see authority, do that) is a catalogue of the same kind as the trade's, and fails for the same reason: the objects differ across persons and moments, and the catalogue cannot be complete. By Proposition XV the large terms are selection and repetition, which are structural: who is reached, how often, through what channel, under what already installed custom. A defence that does not alter those terms, and that instead tries to out-script each script, leaves the causes in place. By Proposition XXII what must be constructed is a contrary affect at exposure. The reliable way to have such an affect available is not to memorise a tactic for each cue. It is to have, as a standing disposition, a small set of contrary objects already paired with the usual exposures: a custom of asking the cause, a reduced access to the channel, a regard that does not depend on the audience that is selling. Those are structural. They work across situations because they do not have to guess the situation. Nevertheless.

∎

Corollary. Reducing repetition is a larger defence than improving detection, because custom is produced by repetition (Definition XI) and detection is an affect that fatigues.

Corollary II. Changing the audience whose regard one lives by is a larger defence than refusing particular shames, because honour is already installed and portable (Proposition XVII).

Scholium. The tactical kit is attractive because it leaves the rest of a life untouched. One keeps the feed, the honour of the group, the saturation, and one adds a skill. The skill is then asked to do the work of the structures it refused to change, and it fails in the ordinary way: at the third hour, in the old mood, with the old audience watching. The structural defence is less attractive because it looks like deprivation. It is not deprivation of influence. It is the refusal to keep paying for the two largest terms, repetition and selection, while complaining about the residual.

A man who does not own a channel cannot be asked to become his own censor in every hour that the channel is open. He can be asked to close it more often, to know which of his affects it selects for, and to have one contrary pairing ready. That is already more than a kit, and less than a monastery. The book will not pretend it is more.

Chapter 65

Proposition XXIVFreedom consists not in exemption from influence, which is impossible, but in the increase of adequate causation in one's own action, which is an increase of action in the sense of Definition III.

Depends on: Definition I, Definition III, Lemma of Part V, Axiom III, Proposition III, Proposition IV, Proposition XI, Proposition XII, Proposition XIII, Proposition XIX, Proposition XXI, E3P1.

Proof. By Axiom III and Definition I no mode is exempt from being caused, and influence is a name for causation so far as it alters the power of acting. Exemption is therefore impossible, and the desire for it is a confused idea of freedom. By the Lemma of Part V and Proposition XXI the freedom available to a man is the increase of the proportion in which he is the adequate cause of what he does. By Definition III and E3P1 that increase is an increase of action, and a decrease of passion, in the technical sense. By Proposition III a decrease of the inadequacy of the idea of the cause is a decrease of the manipulation-relation. That decrease is a necessary condition of the increase of adequate causation. It is not identical with it. A man may have a clearer idea of an external cause and still not be the adequate cause of the act. The further step is the Lemma of Part V: freedom is the increase of the proportion in which he is that cause. By Proposition IV and Proposition XIX the means is causal knowledge joined, by Proposition XII, to a contrary affect. By Proposition XI the operations to be known are few. By Proposition XIII they are ordinary. The demonstration does not identify adequate causation with power to act in every sense. A gain in the power to act that leaves the ideas of one's actions inadequate is an increase of passion, not of freedom, as Corollary II will say. Nevertheless.

∎

Corollary. A gain in the power to refuse one influence, purchased by a loss in the power to act (by exhaustion, isolation, or a governing fear of being moved), is not a gain in freedom.

Corollary II. A gain in the power to act that leaves the ideas of one's actions inadequate is an increase of passion, not of freedom, even when it is successful in the world's terms.

Scholium. This is the last Proposition because it is the first claim of the book, now derived rather than announced. The literature of influence, on both its selling side and its defending side, agreed on a false picture: a force that invades, and a citadel that might hold. There is no invasion, because there is no inside exempt from causes. There is no citadel, because a mode is not a kingdom. There is a proportion. The proportion can be worked at. It cannot be completed. The man who claims to have completed it is selling the citadel.

What the derivation licenses is modest and usable. Ask what affect was already present. Ask what condition changed or what contrary cause was removed. Ask who repeated the image and who owned the channel. Ask whether the action can be understood through an affect and a condition you can name. If you can name no cause, you are, to that extent, manipulated, whether or not anyone intended it. Naming a cause is not yet adequacy. If the named cause is an affect already present and a condition that changed, and if a contrary affect is available to displace what you do not wish to continue, you have done the only thing the axioms permit. The rest is noise, merchandise, and the honour of having a theory of one's own importance.

Finis.

Chapter 66

The Cause Question

Each protocol is a way of forming the idea that Proposition XIX requires, or of constructing the affect that Proposition XXII requires. None is a technique of influence. None is a guarantee. Each cites the Proposition that authorises it.

From Propositions III, IV, and XIX.

When an action is imminent and its urgency is not the urgency of a body in immediate danger, ask: what is the cause of this action, stated as an affect and a condition, not as a justification? If the answer is a property of the object (it is good, it is necessary, it is what people do), the idea is still inadequate. If the answer names a fear, a hope, an honour, a pairing, a repetition, or a hunger that was already there, the idea has begun to be adequate. The question is asked of oneself. Another person cannot answer it for you, and a seller will answer it incorrectly by interest (Proposition VIII).

A worked case. A man is about to send money to a campaign that has written to him every day for a month. The justification is that the hour is late and the cause is just. The affect already present is a fear of a future in which his side has lost, and an honour attached to being the sort of person who did not stand aside. The condition that changed is the daily letter, which refreshed the image. The cause-question answer is not "because they are right." It is: fear of an uncertain political future, honour before an imagined audience of the faithful, and a pairing installed by thirty repetitions. Whether he then sends the money is a separate fact, and will depend on whether a contrary affect is available. The question has already done its work if he can say this. If he cannot, he is, to that extent, manipulated, including by a campaign he loves.

The question has a failure mode, which is the production of a clever story that still conceals the affect. "I am supporting competence" can be the honour of appearing competent. "I am being careful" can be the fear of a later shame. The test is whether the named cause would still move him if the audience were removed and the repetitions stopped. If it would not, the audience and the repetitions were the larger terms.

Chapter 67

The Affected-Object Audit

From Propositions II, XI, and XIV.

Name the object that is actually moving you, which is often not the object being shown. A campaign shows a product and moves a fear of falling behind. A feed shows a face and moves the honour of being current. A warning shows a statistic and moves an old humiliation. Write the shown object and the moving object as two names. If they differ, selection has occurred and the technique, if any, was the attachment. The attachment is ordinary. The concealment of the moving object is the inadequacy.

Chapter 68

The Distribution Audit

From Propositions VII and XV.

Ask who controlled the channel, the repetition, and the alternatives. If you cannot name these, you do not yet have a cause, only a scene. A message encountered once in a conversation and a message encountered forty times in a owned feed are not the same influence. The audit is especially useful against memoirs of genius, including one's own. Genius does not repeat itself forty times. Budgets do.

Chapter 69

The Repeat-Exposure Rule

From Definition XI and Proposition XXIII.

Custom is produced by pairing. The structural defence is to interrupt the pairing, which means reducing the repetition before trying to out-argue the hundredth instance. This is not a purity rule and not a retreat from the world. It is the recognition that Axiom VII is not a fair fight against a channel that never sleeps. Close the channel more often than you debate it. The debate, if it is needed, is cheaper when the pairing has not been refreshed.

Chapter 70

The Approval Ledger

From Proposition XVII and Corollary II of Proposition XXIII.

Keep, for a short time, a record of acts done for a regard you cannot state the reasons of. The record is not a confession. It is a way of making the cheapest lever visible as a cause. If the same audience appears beside many acts, that audience is a structural term, and changing it (or changing your dependence on it) is a larger defence than refusing the next particular shame. Empty honour wanders with popular approval. A ledger shows the wandering.

Chapter 71

The Argument Abandonment Rule

From Axiom VII, E4P14, and Proposition XXII.

When a demonstration has been understood and the affect has not moved, stop adding demonstrations. Further argument will be taken as a new object for the same passion, or as a slight. The available act is the construction of a contrary affect, or the postponement of the encounter until one is available, or the acceptance that this affect, today, is the stronger. Abandoning the argument is not a concession that the argument was false. It is obedience to the axiom that made the book possible.

A worked case, which is also the book's own limit. A reader who has followed the demonstrations and is still governed, in the next hour, by a feed, an honour, or a fear, has not refuted the book. He has confirmed E4P14. The next page will not help him. A walk, a meal, a closed channel, a conversation with a person who can name a cause, a night of sleep in which the pairing is not refreshed: these are not recommended as wellness. They are the available contrary affections, or the conditions under which one might arise. The book that continued past this point, adding proofs to a passion, would have become an instance of the thing it described.

Chapter 72

A day, stated as causes

The protocols are easier to caricature than to use, because they sound like a programme and are in fact a handful of questions. A day in which they are used does not look like a conversion. It looks like this.

Morning: the first images are of other people's regard and of uncertain futures. The Repeat-Exposure Rule is the only one that matters at that hour, because the pairing has not yet been refreshed and is cheapest to interrupt. The channel stays closed until an act that is already determined (work, a child, a meal) has begun from a cause that can be stated.

Midday: a message arrives that uses fear and honour stacked, as in the retirement advertisement of Part V. The Cause Question and the Affected-Object Audit take two minutes. The moving object is named. If it is an old fear, the Distribution Audit asks how often this image has arrived this week. If the answer is many times, the structural term is the larger one, and arguing with the copy is the wrong spend.

Evening: the Approval Ledger, if it is being kept, shows three acts done for a regard whose reasons cannot be stated. That is information about a structural term, not a verdict on the day. The Argument Abandonment Rule applies to any dispute that has already been understood and has not moved. Sleep is, among other things, an interruption of repetition.

Nothing in this day is exemption. The man is moved by hunger, by the work, by the child, by whatever honour remains. He is, in a larger proportion than the previous day, able to name the moving. That is the only increment the last Proposition licenses. It is also the only increment that cannot be sold back to him as a kit.

Chapter 73

What the protocols are not

They are not a method for moving others. The book does not contain the inverse of its own argument, and the reader who came for that inverse has been answered since Proposition I.

They are not a political programme. A state that used these questions on its own propaganda would be a different state, and this treatise does not derive that state. Spinoza's political writings are the place for that work, and they are not repeated here.

They are not a claim that the author is the adequate cause of his own book. The book is determined by the same order it describes: by a literature that needed answering, by a set of documents, by a temperament already disposed to distrust the unmoved mover. Axiom V applies at home. The test of the book is whether a reader can state more of the causes of his own action after it than before. The test is not whether he admires the author. Admiration is honour, and honour is the cheapest lever.

Chapter 74

A Closing Note

The book began with a confusion of tongues and ends with a small set of questions. That is the correct proportion. The subject was never as large as the trade needed it to be. Influence is causation under another name. Manipulation is a privation of knowledge about the cause of one's own action. Propaganda is organized repetition addressed to a multitude. Freedom is the increase of adequate causation. None of these is a mystery, and none of them requires a hidden king or a secret kit.

The remaining work is not literary. It is the ordinary work of noticing what is already moving, interrupting a pairing, and pointing a stronger affect at a truer object. A man who does that more often than he did last year has used the book. A man who can recite its Propositions and is no more the adequate cause of his actions than before has acquired another honour. The book has nothing further to sell him.

Chapter 75

Appendix A. The Replication Ledger

The appendices collect, in one place, the empirical ledger, the documentary chronology, the deductive chain, the vocabulary, and the tactics of the trade sorted by what can be said of them. They add no new demonstrations.

The figures below are taken from the evidence audit compiled for this treatise. Four headline numbers were checked independently against publisher metadata: Tannenbaum 2015, Wagenmakers 2016, Vohs 2021, and Coppock, Hill and Vavreck 2020. Four further figures (Hagger 2016; Open Science Collaboration 2015; Sethuraman 2011; DellaVigna and Linos 2022) were read from the primary PDFs by the compiler of the audit. Those four are marked compiler-read wherever they appear as load-bearing numbers. Where the audit records uncertainty, the table records it.

Chapter 76

Appendix B. Bernays Chronology

Contested entries are marked. Sources are those of the documentary dossier.

Chapter 77

Appendix C. Table of Propositions

Chapter 78

Appendix D. Glossary

Spinoza's terms, as used here.

Affect. A modification of the body by which its power of acting is increased or diminished, and the idea of that modification. Not a mood in the ordinary sense, and not a feeling opposed to reason.

Adequate cause. A cause through which the effect can be clearly and distinctly perceived (E3D1).

Conatus. The endeavour of each thing to persist in its being, which is its actual essence (E3P6, E3P7).

Custom. A disposition produced by repeated pairing of affections (E3P14; Definition XI).

Freedom. As Spinoza, the thing that exists and acts from itself alone (E1D7; Definition XII). For a finite mode, the increase of adequate causation in its own action (Lemma of Part V). Not exemption.

Inadequate idea. An idea that involves a privation of knowledge (E2P35). It follows by the same necessity as an adequate idea (E2P36).

Multitude. Men considered so far as they are determined by affects rather than by reason (Definition X).

The trade's terms, defined and then set aside.

Dark psychology. A commercial name for ordinary operations on fear, honour, and ignorance. It names a market, not a mechanism.

Frame. A description under which an already present desire is attached to an object. Not a force.

Nudge. A change of condition, usually a default. Real, small after bias correction, and not a theory of mind.

Priming. In the large social form, a failed claim. In the narrow automaticity form, a small residual that does not license the catalogues.

Technique. A named form of words, gesture, or script offered as a portable cause of influence, considered apart from the sufficient set. What it can do, when it does anything, is change a condition or remove a countervailing cause (Proposition I). As a transferable secret, nothing.

The unconscious, as used by the trade. A hidden room in which a manipulator works. Spinoza's account is the opposite: there is no such room, only causes of which we are ignorant, including the manipulator's own.

Chapter 79

Appendix E. The Tactics Catalogue, Sorted

Real and measured. Anchoring (shifts a number). Gain-loss framing (replicates smaller than first reported). Mere exposure. Defaults, with large heterogeneity. Fear appeals of modest average size, stronger with an efficacy statement. Asch-type conformity to a unanimous majority. Reciprocity when the gift has cost. Scarcity so far as it is reactance or a popularity signal.

Real but trivial or narrow. Nudges after publication-bias correction. Advertising's short-term elasticity. Political advertising's average effect on vote. Subliminal cues among the already prepared. Facial feedback as a small, method-dependent residue. Cialdini's liking.

Unmeasured as science. The 48 Laws. Dark psychology kits. "How to read anyone." Most NLP packaging beyond the already-separate rapport literature. Memoir claims of having won an election or invented a want.

Refuted or hoax. Vicary's cinema. Bargh-style social priming as sold. Money priming. NLP eye-accessing cues. Mehrabian as a general law of communication. Stanford Prison as a demonstration of roles. Cambridge Analytica as an engine of elections. Power posing as a hormonal and risk intervention.

The sorting is the book's argument in another form. What is real is a change of condition, a repetition, or an affect already present. What is sold is a technique that does not need those things.