On Power

The people who openly seek power are rarely the ones who hold it longest. The people who hold it longest are usually the ones who convinced everyone else that they were not seeking it at all. That is not cynicism. That is the first thing you need to understand about how power actually works, as opposed to how it is described. There is a mode of thinking that treats power not as a possession but as a relation, and almost everyone who has not internalised this gets the basic unit of analysis wrong. The common error is to imagine power as a substance: something a person or institution has, like money in a vault, measured by titles, armies or net worth. The person who has internalised this mode of thinking instead asks, in any situation, who needs what from whom, who can withhold it and at what cost, what each party would do if the other’s cooperation disappeared tomorrow and what each party currently believes about the other that may or may not be true. The central insight that unlocks most of the field’s value is this: power is the structure of dependency between parties, and perception is the medium through which that structure is built, concealed and exploited. Almost everything else, authority, coercion, influence, legitimacy, charisma, reputation, is a special case or a disguise of dependency operating through perception. Once this is seen, hierarchies stop looking like fixed facts and start looking like temporary settlements that hold only as long as the dependencies and perceptions that sustain them remain unchanged, and most of what looks like inexplicable institutional behaviour, irrational decisions, sudden reversals, the survival of the mediocre and the destruction of the capable, becomes legible. The title under which this field is often discussed makes people uncomfortable, and that discomfort is itself the first thing worth examining. There is a widespread conditioning to believe that power, manipulation, strategic deception and the study of how these forces operate are subjects fit only for those who intend to misuse them, and that the ethical person has no need of this knowledge. This belief is not only mistaken. It is dangerous, because it leaves the person who holds it at a permanent disadvantage in every environment where dependency, coalition and perception determine outcomes, which is every environment that matters. The Anatomy of Dependency: A and B Revisited The foundational vocabulary of the field begins with Robert Dahl’s deceptively simple formulation: A has power over B to the extent that A can get B to do something B would not otherwise do. This definition, often dismissed as too thin, is in fact the load-bearing floor on which everything else is built, because it forces attention onto behaviour and counterfactuals rather than onto titles or reputations. The crucial move is to ask what B would have done in the absence of A’s intervention, and this counterfactual question is where most popular thinking about power goes wrong, because it is far easier to observe who gives orders than to observe what would have happened without them. Steven Lukes extended this in Power: A Radical View by distinguishing three faces of power. The first face is the Dahlian one: direct behavioural influence in observable conflicts. The second face, drawn from Bachrach and Baratz’s work on agenda setting, is the power to keep issues off the table altogether, so that B never even raises a demand A would have had to resist. The third face, the most demanding, is the power to shape B’s preferences and perceptions so that B does not want the thing that conflicts with A’s interests in the first place. This third face is why dependency, not behaviour, is the right unit of analysis: a person whose preferences have been shaped need never be coerced, because coercion only becomes visible at the point where dependency is contested, and a sufficiently complete shaping of preferences removes the contest before it begins. Max Weber’s distinction between Macht (Power) and Herrschaft (Domination/Authority) are primarily outlined in his posthumously published magnum opus, Wirtschaft und Gesellschaft (Economy and Society). The difference maps onto this same structure from another angle. Macht is raw capacity to impose one’s will even against resistance; Herrschaft is the condition where commands are obeyed as a matter of course, without the need for resistance to be overcome at all, because the relation has been institutionalised into something that looks like legitimate authority. Weber’s three types of legitimate authority, traditional, charismatic and rational legal, are best understood not as three different kinds of power but as three different stories that dependent parties tell themselves about why their dependency is acceptable. The story changes; the underlying dependency structure is what actually has to be tracked. Jeffrey Pfeffer’s organisational research and Robert Caro’s biography of Robert Moses both confirm, from opposite ends of the literature, that authority on an organisational chart and power in this dependency sense routinely diverge, and that the gap between the two is where the most consequential dynamics in any institution actually take place. The Currency Problem: Why Resources Are Not Power A persistent error is to confuse resources with power itself. Resources, money, weapons, votes, information, formal authority, skill, are potential power, convertible into actual power only through a process that can fail at multiple points. This is the paradox of unrealised power: an actor with overwhelming resources can lose every confrontation because those resources cannot be converted into leverage over the specific dependency at stake. The clearest illustration is the BATNA framework from Getting to Yes by Roger Fisher and William Ury. Bargaining power in a negotiation is not a function of overall wealth but of one’s best alternative to a negotiated agreement relative to the other side’s. A wealthier party with no good alternative can be out-leveraged by a poorer party who can credibly walk away, because power in that moment is defined by the relative cost of the relationship ending. Thomas Schelling’s The Strategy of Conflict pushes this further: power often resides not in the ability to act but in the ability to credibly commit to a future action, including one that would be costly to oneself. A threat too costly for the threatener to carry out is no threat at all unless the threatener can bind their own future behaviour, which is why Schelling treats the deliberate reduction of one’s own options, burning bridges, making promises public, delegating to inflexible agents, as a source of power rather than a loss of it. This inverts the naive equation of power with freedom of action. Often the actor who has visibly removed their own options has thereby made their commitment credible and gained leverage over a party who must now adjust to that fixed point. Caro’s account of Moses’s bond covenants is the institutional version of this logic: by writing the revenue structure of his public authorities so that no elected official could redirect their funds without enormous political cost, Moses converted a constraint on himself into a source of power that outlasted every governor who tried to remove him. Expertise is a clear case of the conversion problem and deserves separate treatment. The typology of power bases developed by John French and Bertram Raven distinguishes expert power, knowledge others depend on, from legitimate, reward, coercive and referent power, which derive from position, control over outcomes or likeability. Jeffrey Pfeffer’s Power: Why Some People Have It and Others Don’t develops the relevant implication: expertise converts into power only when it is both depended upon and difficult to replace, and neither condition can be assumed. A specialist whose knowledge is widely shared or easily documented has no leverage regardless of its depth, because the dependency it might create can be satisfied elsewhere at low cost. Pfeffer’s further point is that competence is no substitute for the coalition and positioning work covered later. The skilled person who assumes their competence will be recognised on its own terms makes the same error as the person who mistakes resources for power in general: skill is potential power and, like any resource, requires dependency and irreplaceability to become actual power. The Hidden Curriculum of Hierarchy: Greene, Gracian and the Court The popular literature on power, most visible in Robert Greene’s The 48 Laws of Power, is often dismissed as cynical, but it earns its place because it documents, with unusual precision, how the second and third faces of power operate at the interpersonal scale inside hierarchies. Greene’s emphasis on never outshining the master, on saying less than necessary and on concealing one’s intentions, is, stripped of its packaging, a description of how dependency is managed and concealed within courts, corporations and bureaucracies. Never outshining the master describes a predictable structural dynamic rather than an occasional pathology: authority that rests on insecurity does not reward capability that makes it feel threatened, it eliminates that capability, and the practical response is to frame one’s competence as an extension of the superior’s position until accumulated influence makes that framing unnecessary. Saying less than necessary and concealing intentions both reduce the information leaked about one’s thinking, and since every word spoken is data others use to model and anticipate the speaker, controlling that flow is itself a form of power, the mirror image of the information control Caro documents in Moses. The throughline of this literature, back to Baltasar Gracian’s The Art of Worldly Wisdom and Niccolo Machiavelli’s The Prince, is that visible power is fragile because visibility invites coalition formation against it. Machiavelli’s central insight, often flattened into a slogan about ends justifying means, concerns timing and perception: the prince who is feared but not hated, who appears virtuous without being constrained by virtue, has converted the appearance of legitimacy into a resource that reduces the cost of maintaining dependency. Gracian’s maxims compress this into individual practice: never let others see the bottom of you, because a fully knowable person is fully predictable, and always keep something in reserve, because an actor who has visibly committed all their resources has nothing left when the situation changes. This is the same mechanism Weber describes as the routinisation of charisma into traditional or legal authority, seen from inside by a practitioner rather than from outside by a sociologist. The Coalition Engine: Why Decisions Look Irrational An account of power that stops at the interpersonal scale cannot explain why entire institutions, including democracies, behave in ways that look irrational against their stated goals. The coalition theory of Bruce Bueno de Mesquita and Alastair Smith, set out in The Dictator’s Handbook, supplies the mechanism. Every leader depends on a winning coalition large enough to keep them in place, and the size of that coalition relative to the pool of people who could plausibly belong to it predicts more about institutional behaviour than any constitution or mission statement. Where the coalition is small relative to the pool, each member’s loyalty is cheap because each member is replaceable, and the leader rewards loyalty privately while neglecting everyone outside the coalition: the structural signature of kleptocracy. Where the coalition is large, private rewards cannot be stretched far enough, so the leader must deliver public goods that benefit even those outside the coalition: the structural signature of governance that looks, from outside, like a commitment to the common good, when it is a coalition too expensive to bribe individually. This also explains the persistent pathology of large coalition systems: the constant pressure to shrink the effective coalition through gerrymandering, the manipulation of information environments and the cultivation of tribal loyalties that override policy assessment, because a smaller coalition is always cheaper to maintain. A policy superior by every measurable standard can fail repeatedly for reasons that have nothing to do with its merits and everything to do with whose loyalty payments it disrupts, which is why coalition mapping, identifying who must be satisfied and what each requires, is as much a part of understanding power as identifying the dependency itself. The Psychological Substrate: What People Are Actually Responding To Dependency structures are activated and maintained through specific features of human psychology, without which the field is incomplete. Robert Cialdini’s six principles of influence, reciprocity, commitment and consistency, social proof, authority, liking and scarcity, set out in Influence, describe mechanisms that evolved because in most circumstances they are reliable shortcuts to good decisions: a favour usually indicates the giver values the relationship, and a credential usually indicates competence. The difficulty is that all six can be triggered artificially with much the same effect as their genuine counterparts, so a strategically timed gift produces real obligation even when its purpose was instrumental, and manufactured social proof produces real conformity pressure even when the appearance of consensus was constructed. Greene’s The Laws of Human Nature extends this into dispositional psychology, arguing that the forces most consequential for power are the ones people least admit to in themselves. Envy is the clearest case: almost never acknowledged as envy, rationalised instead as principled concern, yet the pattern of who attracts disproportionate opposition tracks relative standing more reliably than it tracks any objective failing. Subclinical narcissism is overrepresented in leadership positions for the same reason: the need for one’s vision to be implemented without modification, and the experience of disagreement as disloyalty, are also the traits that drive people toward positions where disagreement can be minimised. This is not a separate subject from dependency; it is the mechanism by which Lukes’s third face is implemented at the level of individual interaction. Cialdini’s Pre-Suasion adds the temporal dimension. The six principles govern an influence attempt itself; Pre-Suasion governs the moment before, the management of the psychological state in which a message is received. Whatever a person attends to immediately before a request arrives primes their thinking in ways that are specific and largely unconscious: a person primed to recall a success responds more favourably to a proposal, a person surrounded by images of achievement sets higher goals than the same person in a neutral setting. The agenda of a meeting, the setting of a negotiation, the small talk before a difficult conversation are not incidental, they are engineered context, and the practitioner who understands this designs context and message as a single intervention. Reading the Room: Perception Before Position Everything above assumes an actor can accurately perceive the dependency structure, the coalition and the psychological state of others, but that perception is itself a skill with its own literature, and it sits upstream of every strategic move described elsewhere. Joe Navarro’s What Every Body Is Saying rests on the distinction between the verbal channel, conscious and controlled, and the nonverbal channel, posture, gesture, microexpression, the orientation of feet and torso, the timing of eye contact, which is largely involuntary because it originates in the limbic system. The limbic brain produces freeze, flight or fight responses faster than the conscious mind can intervene, which is why a person who feels threatened by a question will exhibit a momentary stilling of movement before answering, even when the answer itself is fluent. The discrepancy between the controlled verbal channel and the uncontrolled nonverbal one is not proof of deception in any simple sense, but it is one of the most reliable signals that there is more in an area than is being said, and an actor who can read this discrepancy has information about the dependency structure that the other party does not know they have revealed. Gavin de Becker’s The Gift of Fear addresses the same perceptual layer from the angle of self-protection, but the claim generalises. De Becker argues that fear, as distinct from social anxiety or the discomfort of seeming rude, is produced by a threat detection system usually more accurate than conscious reasoning, and that most people override it because acting on it risks seeming paranoid. His catalogue of pre-incident indicators, charm deployed where charm is unusual, unsolicited favours that manufacture obligation, forced teaming that creates artificial intimacy and lowers defences, and typecasting, a mild insult designed to provoke a defensive response that creates an opening, describes tactics that recur not only before physical violence but before manipulation in professional and political settings, where they function as opening moves in establishing the kind of asymmetric dependency the rest of this essay is about. The dependency structures, coalitions and psychological triggers described in the preceding sections do not announce themselves; they are built through small interactions whose significance is legible mainly through the channels Navarro and de Becker describe and an actor untrained in this perception is working from a systematically incomplete map regardless of how sophisticated their strategic framework is. The Limits of Confrontation: Strategy Before the Fight The frameworks above describe the structure of power at a given moment, but Sun Tzu’s claim in The Art of War, that supreme excellence consists in subduing the enemy without fighting, concerns the sequence through which that structure is altered before confrontation becomes necessary. Every confrontation, even a won one, consumes resources, relationships and goodwill, whereas a position constructed so thoroughly in advance that resistance appears futile costs nothing at the point of use. This reframes mechanisms already described: controlling the agenda is not merely a way of avoiding a contest, it is the specific technique by which Sun Tzu’s principle is implemented in institutional life, and information asymmetry, knowing more about the other side’s intentions and constraints while revealing less of one’s own, is the resource that makes pre-emption possible, which is why Sun Tzu treats deception as a strategic instrument rather than a moral failing. Robert Greene’s The 33 Strategies of War adds the corollary failure mode: the person who recognises a developing conflict only once it becomes openly confrontational has already ceded the initiative, because the early signals, a withdrawal of support, a change in who is consulted, were available earlier and went unread, often because the perceptual training described above was absent. Miyamoto Musashi’s The Book of Five Rings supplies the internal counterpart: a mind attached to a particular outcome telegraphs that attachment, and telegraphed need is itself a dependency an adversary can exploit, so the capacity to pursue an objective without being psychologically dependent on its attainment, what Musashi calls the void, is a precondition for strategy rather than a virtue layered on top of it. Fearlessness as Foundation Almost every form of leverage described above operates by making someone afraid of something, of losing support, missing approval, suffering retaliation or facing the alternative the other party represents. The 50th Law, built around the trajectory of 50 Cent from South Jamaica through repeated attempts on his life to global commercial influence, argues that an actor who has confronted the worst realistic outcome available to them and found it survivable has removed the primary handle by which others exert control, because the standard mechanisms of control route through loss aversion and a person not governed by loss aversion is not governed by them. This connects directly to Musashi’s void: the strategist not desperate to avoid a particular outcome perceives the situation more accurately, because fear narrows attention to the threat and away from the wider field in which alternatives also exist. The implied exercise is one of accounting rather than performance: what is the worst realistic outcome of the current course, and what would the actor’s situation actually look like if it occurred. In most cases the answer is that the feared outcome is less severe than the fear response suggests and that the actions taken to avoid it cost more, in autonomy and in the quality of decisions made under that fear, than the outcome itself would have. The Dark Attractor: What Power Does to People The literature most oriented toward acquiring power consistently understates what sustained power does to the person who holds it, and how the dependency structures described throughout this essay change once the person at their centre has been altered by holding them. Caro documents this with precision in the arc of Moses from the 1920s to the 1950s. Philip Zimbardo’s The Lucifer Effect, built around the Stanford Prison Experiment, identifies the situational forces that produce this deformation: randomly assigned student volunteers became abusive or broken within days, because role definition, gradual escalation, social proof, the creation of in-groups and out-groups and the removal of individual accountability are sufficient to override individual character in most people most of the time. Hannah Arendt’s Eichmann in Jerusalem shows what happens when this trajectory completes: the substitution of procedural correctness for independent judgment, reached not through a single decision but through an accumulation of individually reasonable steps. The pattern is consistent. Power begins as an instrument for purposes the person values. The pursuit of power gradually becomes the primary activity, displacing the purposes that motivated it. Its exercise begins to feel intrinsically rewarding, challenges to it feel existential and the people over whom it is exercised cease to register as people with independent interests. Ryan Holiday’s Ego Is the Enemy names the mechanism at the centre: the defensive need to protect a specific image of oneself. The ego-driven actor cannot take accurate feedback because feedback threatens self-image, cannot acknowledge error because error is too costly to the self-concept, and cannot read situations clearly because incoming information is filtered through what it means about the self rather than what it means about the situation, the same distortion Musashi and the 50th Law identify as the strategist’s primary liability. This is what happens to the perceptual and strategic apparatus described in every preceding section once the actor operating it has been altered by the dependencies they were managing. The trajectory is not inevitable, but it is the default without specific resistance, and the literature points to what that resistance consists of: an identity prior to and independent of one’s institutional role, so the role can be used rather than defended as the self; relationships outside the sphere of one’s power, the only source of unmanaged feedback that power systematically filters out; and the recurring practice of independent judgment as a discipline, since Arendt’s account shows its erosion is gradual and invisible from inside. The Unified Picture Every idea above expresses the same claim: power is the structure of who needs what from whom and at what cost, perception is the medium through which that structure is built, concealed and exploited, and the forms power takes, titles, charisma, network position, psychological influence, coalition arithmetic, are mechanisms through which dependency and perception interact. Dahl’s definition identifies the structure at its most visible. Lukes’s second and third faces show how it operates by shaping which dependencies are perceived, so the most stable forms of power never need to be exercised. Cialdini’s principles, Pre-Suasion and Greene’s account of envy and narcissism show the psychological machinery through which that shaping is carried out, while Navarro and de Becker show how its early signals can be read before the dependency becomes visible any other way. Bueno de Mesquita and Smith’s coalition arithmetic shows the same shaping at the institutional level, where coalition size determines whether power is exercised through private reward or public good. Schelling’s commitment logic, the BATNA framework and Sun Tzu’s pre-emption show the structure can be deliberately engineered, while Musashi and the 50th Law show this engineering depends on a perceptual clarity that fear and ego both degrade. The failure modes, dependency blindness, principal agent capture, the iron law of oligarchy and the corrosion documented by Caro, Zimbardo and Arendt, are cases where an actor stops tracking the dependency structure, either by mistaking a resource for power, by allowing delegation or coalition drift to create unmonitored dependency, or because the actor’s own perceptual apparatus has been degraded by the power they hold. The 20 percent that produces 80 percent of the field’s power is therefore the counterfactual definition of dependency from Dahl, the extension of that dependency into perception via Lukes and the psychology of influence, the recognition from Schelling and Sun Tzu that dependency and the timing of its contestation can be engineered and the recognition from Musashi, the 50th Law, Holiday and Arendt that the actor’s own apparatus is part of the system and not exempt from its dynamics. The Limits The dependency framework’s own blind spot is that it can become a universal solvent explaining everything and therefore predicting nothing, since almost any outcome can be redescribed after the fact as evidence of some hidden dependency or coalition calculation with no independent way to have predicted it beforehand. A second limit is that this emphasis, taken to its extreme, produces a permanently adversarial reading of every relationship, including ones better described as cooperation or trust. Not every asymmetry is exploitative and not every gift creates a debt; a frame that treats every relation as a covert negotiation, and every interaction as an opportunity to read or be read, can become a blindness of its own, corroding the relationships the closing section identifies as the antidote to power’s corrosive effects. Third, the perceptual tools in the Navarro and de Becker section carry a specific risk: nonverbal signals and intuitive threat detection are probabilistic and context dependent, frequently misread by people who have learned enough to feel confident without having learned enough to be accurate, and constant reading of others’ internal states can produce a paranoia that degrades judgment in exactly the way the framework warns against. Finally, practitioners tend to overestimate how legible and controllable dependency structures are. Schelling’s commitments, Sun Tzu’s positioning and Greene’s tactics read as though the social world were a board on which a skilled actor can engineer outcomes, but real dependency structures are emergent and frequently invisible even to embedded participants until they fail, and the same blindness the framework diagnoses in others is available to anyone applying it with too much confidence in their own clarity, which is the ego distortion Holiday and Musashi identify as the strategist’s central liability. What This Demands of You You must ask, in any relationship or institution you encounter, what each party would do if the other’s cooperation disappeared tomorrow, because that counterfactual is the actual content of the power relation, regardless of titles or appearances. You must treat resources, including your own skills, as raw material that may or may not convert into leverage, and ask what the conversion mechanism is and whether you are difficult to replace, rather than assuming possession or competence equals power. You must notice when an issue never gets raised at all and ask who benefits from its absence from the agenda. You must map the coalition any decision-maker depends on before assuming a policy’s merits will determine its fate, and ask what each member would lose if it were adopted. You must notice when your own reactions, gratitude, conformity, deference to credentials, sudden intimacy with a near stranger, have been shaped by an engineered context, and ask whether the reaction would survive being made explicit to you. You must read the early, often nonverbal, signals of developing conflict rather than waiting until confrontation is unavoidable, while holding those readings as probabilistic rather than certainties that justify pre-emptive hostility. You must notice when your own need for an outcome is visible to others, since that visibility is itself a dependency handed over, and you must do the accounting of what the worst realistic outcome of your current position actually is, since most of the fear that distorts perception turns out, on inspection, disproportionate to what it protects against. And you must, repeatedly rather than once, ask what sustained possession of whatever power you hold or seek would do to your judgment, your relationships and your sense of what the power is for, with people in your life who have no stake in the answer you would prefer. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com