Thomas C. Schelling — The Strategy of Conflict

288 published claims from this book.

• A biased mediator can be accepted when clue-rich situations block focal points
• A clever tactic can be defeated when adversaries study and counter it.
• A conditional commitment fails if its conditions cannot be verified after the fact
• A credible conditional commitment must combine both a promise and a threat.
• A credible threat can subordinate the final decision to chance outside the threatener’s control
• A credible threat works precisely because it need never be carried out.
• A deterrence threat often hurts the threatener as much as the threatened
• A deterrent threat is only credible if control over its triggering is partly surrendered.
• A deterrent threat is only too large when its size damages credibility
• A dominant minority can hold power through threats and enforced coordination costs
• A fractional commitment blends a threat with a promise
• A fulfilled threat gains credibility from its value to future threats, not present gain
• A losing side rationally abandons containment for deliberate risk-creation.
• A mediator can compare offers without revealing them
• A mediator can create focal points through dramatic suggestion
• A mixed-strategy equilibrium can be jointly dominated by a pure coordinating point.
• A normative theory must exploit labels, not expunge them
• A perfectly stable balance of deterrence may not even be desirable, let alone achievable.
• A probabilistic threat stays credible despite failures because one trial proves little
• A promise must be enforceable or releasable by the receiving party to work
• A randomized minimax strategy deliberately destroys all communication between players
• A third party can provoke war between two others by staging an ambiguous act
• A threat backed by reputation cannot be unilaterally renounced by its target
• A threat binds only if the other believes it cannot be retracted
• A threat must be credible and credible threats require commitment
• A threat only coerces when the threatener has visibly left himself no choice
• A threat to act is only credible if one is committed, not merely willing
• Abolishing English law in one province threatens the same arbitrary rule everywhere.
• Abstract theory is clarified only through concrete applied cases
• Addiction compromises are unstable so total abstinence is the stable focal point
• Adding free speech barely changes a tacit bargaining game’s character
• Adversaries in indeterminate games mutually construct the norms that govern their conflict
• Adversaries may gain by buying each other’s defensive capacity
• Agenda confinement and publicity neutralize extortionate threats in negotiation
• Agreements may require redesigning incentives, not just dividing gains
• Agreements stay enforceable through the shadow of future bargains
• Alliances can serve as devices for surrendering freedom of choice
• Among plausible coordination rules, choosing the greater payoff is the more natural
• An act’s coercive cost has both a direct component and a risk-of-escalation component.
• An enforceable promise requires destroying one’s own power not to comply
• An unchallenged proposal coordinates action as well as an accepted one
• Any coordinated expectation is rational regardless of how the players arrived at it
• Any mutually recognized cue can become the coordination key
• Any theory of strategic coordination has an inherently empirical component
• Arms-control compliance cannot be verified because decision rules are unobservable and breach means war
• Asymmetrical narrow-road conflicts empirically support the coordination-signal principle.
• Asymmetry cannot be proven inconsistent with player rationality
• Automatic mechanisms lower risk by deterring tests of resolve
• Automatic threat mechanisms reduce risk by discouraging nerve tests
• Auxiliary tie rules need not dominate analysis of asymmetric games.
• Bare assertions of non-hostility cannot substitute for authenticable facts
• Bargaining games mix positively and negatively sloped payoff pairs
• Bargaining outcomes rest on convergent reflexive expectations, not on force or merit
• Bargaining situations are imperfect-correlation-of-preferences games
• Bargaining uses culturally available materials to build shared traditions of restraint
• Bargains over unobservable compliance must be indexed to observable proxies
• Being rational, informed, and in control is not always strategic
• Being unreachable during negotiation strengthens one’s bargaining position
• Bidding patterns reveal whether a payoff functions as coordination signal or insurance.
• Binding oneself can paradoxically strengthen one’s bargaining position
• Binding-agreement machinery becomes irrelevant under perfect symmetry
• Both rivals seek a detection advantage while hoping none is found
• Both secrets and certain capabilities are best renounced under mutual suspicion.
• Brinkmanship is placing oneself on an uncontrolled slope of shared risk
• Bundling an extortionate demand onto a legitimate agenda makes the threat effective
• Cities without obvious central sites suppress spontaneous mob congregation.
• Civilized states can learn enforcement devices from the underworld and ancient despotisms
• Coalition formation is where coordination theory and sociology meet
• Commitment buys first move while threat commits to a second-move response strategy
• Commitment converts an indeterminate bargaining situation into a determinate two-move game
• Commitment includes removing options, delegating fulfilment, and self-handicapping
• Commitment strategies require communication and risk provoking stalemate
• Commitments are often manufactured by removing one’s own control over action
• Commitments must be communicated through evidence, not just words.
• Commitments, promises, and threats all reduce a player’s own payoffs visibly and irreversibly
• Committing to a strategy can yield a better outcome than remaining uncommitted
• Common knowledge of mutual confidence makes a focal point self-sustaining
• Communication effectiveness is an empirical matter exceeding formal minimum standards.
• Compellence works by creating shared uncontrollable risk of catastrophe
• Complementary agreements are self-enforcing through interdependence and focal points
• Concessions require an obvious stopping point to avoid signaling collapse
• Conflict and partnership-distrust situations share one identical game structure
• Conflict settings share common-interest structures across wildly different scales
• Conflict situations are essentially bargaining situations
• Conflicting bargaining theories reveal mathematicians lack a common aesthetic for selection
• Contractual commitments shift negotiation costs rather than truly binding anyone
• Conventions and constraints reduce ambiguity the way language constraints reduce garbles
• Cooperation may not suffice to arrest instability in the arms race.
• Coordinate behavior emerges from conventions and impromptu codes without explicit plans
• Coordination play seeks a meeting of minds while minimax play avoids any
• Coordination requires a common source of suggestion
• Credible commitment requires finding a penalty against one’s own nonperformance
• Credible threats can be conveyed by action rather than verbal pledge
• Credible threats can require revealing information that defeats the threat
• Credible threats must attach penalties to observable acts, not intentions
• Crisis agreements require both a stable solution pattern and verifiable, cheat-proof monitoring
• Critical thresholds on irrational-attack probabilities trigger joint attack.
• Defining agreement requires specifying concrete recording operations
• Defining rationality to include symmetry strips the theory of empirical content
• Delegating decisions to agents makes threats more credible
• Demonstrated risk can succeed precisely because it is risky
• Desperate mutual need makes even poor signals coordinate action
• Destroying communication can be a winning move in coordination games.
• Deterrence can be built as graduated probabilistic trip wires rather than a single switch.
• Deterrence can operate as an unintended byproduct of other actions
• Deterrence logic can justify seeking more horrific retaliatory weapons.
• Deterrence operates on graduated probability, not a binary trip-wire logic
• Deterrence requires mixed conflict and common interest between parties
• Deterrence rests on surviving retaliation capacity, not on existing destructive capacity.
• Deterrence theory is the theory of the skillful nonuse of force
• Deterrent threats can be passive but compellent threats require commitment
• Enforcing one’s own promises requires a supporting legal or information structure
• Establishing identity and a communication line creates new exploitable threat value
• Even a player sure of his own restraint inherits the other’s amplified fears
• Even conflict between adversaries can be modeled as partnership discipline
• Even explicit bargaining relies on tacit signals to coordinate expectations
• Every bargaining game ends in a simultaneous final move made in ignorance
• Explicit bargaining outcomes rest on prior tacit coordination between rational players
• Extortion and threats require a communication channel to the intended victim.
• Extortion succeeds only when bundled onto a legitimate negotiating agenda
• Failed threats create joint incentives to undo commitments that parties cannot jointly undo
• Fear of surprise attack feeds back into raising each side’s own probability of attacking
• Fear of surprise attack produces all-or-none outcomes, not probability multiplication.
• First-mover advantage decides bargaining when commitments are costless and binding
• First-strike advantage converts defensive fear into preemptive violence spirals.
• Five centuries of use have yielded no adequate defense against the bullet
• Focal points require prominence and uniqueness to coordinate action
• Formal agreements can strengthen tacit coordination norms.
• Formal game theory abstracted away the communication and enforcement systems it needs
• Fungibility arguments fail because tradition governs war limits
• Game theory clarifies deterrence problems but remains too abstract to apply directly
• Game theory succeeds on pure conflict but fails on mixed conflict-cooperation
• Gang leaders command obedience through expected mutual punishment coordination
• Genuine ignorance can be strategically advantageous when an opponent recognizes it.
• Giving up defensive weapons is logically identical to exchanging hostages
• Government asset freezes immunized owners against threats against their relatives.
• Governments can enter major wars inadvertently through imperfect decision processes
• Granting new legal rights can weaken rather than strengthen one’s position.
• Hinted threats gain credibility precisely because they are deniable
• Hostages and shared vulnerability render agreements self-enforcing
• Hostile epitomes can describe real events while distorting them beyond recognition
• Hostility becomes part of the communication structure when it distorts perception.
• Human pattern-seeking is designed for cooperation, not for randomness
• In bargaining, seeming weakness can convert into strategic strength.
• In coordination, neither player can outsmart the other without outsmarting himself
• In focal-point bargaining small concessions signal collapse and are rejected
• In mutual-dependence games possessing initiative can be a disadvantage
• Inability to make enforceable promises can shield someone from extortion
• Incremental gameplay permits adversaries to build trust that rapid moves destroy
• Inspection schemes can gather too much information and increase vulnerability
• Inspection systems should be overbuilt for crisis contingencies beyond their agreed routine purpose.
• Institutional reputation makes threats more credible than individual threats
• Institutions and traditions persist because they are conspicuous coordination solutions.
• Insurance and coordination-clue motives can be operationally distinguished by bidding tests.
• Interaction between two minds cannot be modeled by formal deduction alone
• Invulnerability in strategic argument is often proxied by undetectability
• Irreversible commitments force the enemy, not oneself, to be the one who yields.
• Judging mutual invulnerability requires comparing two distinct deterrent threats
• Known capability for failure can stabilize mutual-threat situations
• Leaving a final standing offer is a winning commitment tactic.
• Limited war deters partly by exposing both sides to uncontrollable risk of escalation
• Limited war’s main function is signaling escalation risk, not local military gain.
• Limited-rationality partners need the solution communicated to them before they can commit
• Limits emerge through tacit jockeying and need unique focal qualities.
• Linking indivisible projects enables compensation that separate bargaining forbids
• Linking two separate negotiations can enable threats impossible within a single game.
• Literal disarmament of horrific weapons can increase instability rather than peace.
• Local conventional defense shifts the final nuclear decision onto the aggressor
• Lotteries make indivisible objects and obligations divisible and fair
• Mandatory secrecy of the ballot protects voters from intimidation by removing proof.
• Mathematical solutions only focus expectations when players perceive each other as mathematicians
• Mathematical symmetry fails as a bargaining principle under value uncertainty
• Mixed strategies primarily prevent the opponent’s anticipation of one’s choice.
• Mixed-motive games require tacit social coordination absent from zero-sum games
• Most conflicts are variable-sum games with common interests
• Move asymmetry is what makes cooperative game theory interesting.
• Moves carry evidential weight that cheap talk cannot convey
• Moves reveal value systems and commit players where speech cannot
• Mutual deterrence depends on balance stability, not on sheer symmetry.
• Mutual fear can amplify exogenously into runaway escalation via feedback loops.
• Mutual invulnerability raises the credibility of limited retaliation
• Mutual nuclear deterrence is a modern form of hostage exchange.
• Mutually invulnerable retaliation forces neutralize each other’s deterrent role
• Mutually perceived signals explain the pull of psychologically dominant equilibria.
• Natural boundaries provide stable stopping points in military and political conflict
• No government has a fully determinate plan for every foreseeable contingency
• Non-symmetric moves like commitments and threats enrich cooperative games.
• Non-symmetry beats both committed symmetry and committed asymmetry.
• Normal-form representation erases the communication and enforcement structures that make tactics work.
• Nuclear arms negotiations are an open-ended bargaining problem beyond fixed payoff matrices
• Nuclear weapons differ by a tradition of their nonuse
• Once a taboo limit is breached, the inhibition is permanently eroded.
• One player’s promise can worsen the other player’s position
• Parliamentary voting order materially shapes which outcomes can be coerced
• Partial randomization can turn a conflict-addled game into a profitable equilibrium
• Particular encounters acquire symbolic weight because players know reputation is at stake
• Parties able to make binding commitments outbargain those who cannot
• Parties can reach win-win deals by creatively expanding the issues
• Parties cannot choose which coordination signal will govern them
• Perfect move symmetry neutralizes all interesting bargaining moves in advance.
• Players can derive coordination signals from labels and connotations beyond payoff structure
• Players may manufacture ex-post reasons to justify coordinating on a salient equilibrium
• Players win only by drawing a division line both can anticipate identically
• Pre-arranged communication infrastructure raises the odds tacit bargaining succeeds
• Preclusive self-defense can itself provoke the attack it tries to avert
• Prisoner’s dilemma arises when dominant incentives yield jointly worse results
• Promises are needed whenever an agreement leaves incentive to cheat
• Public pledges of reputation can make concessions impossible
• Publicly known reluctance to strike first paradoxically deters the enemy.
• Pure conflict almost never occurs even in war
• Pure symmetry can single out a unique tacit-coordination point
• Quota systems replace tacit coordination games with explicit enforced agreements.
• Racketeers enforce cartel agreements that lie beyond legal sanction
• Rational actors need not calculate beyond first-order attack likelihoods.
• Rational agreement operates through shared psychic expectations, not mathematical reasoning
• Rational players can correlate strategies without communicating, even with conflicting preferences
• Rational players cannot opt out of social coordination when efficiency requires it
• Rational players must factor non-mathematical focal details into strategy
• Rational self-defense adjustments are compatible with both sides still preferring mutual non-attack
• Reformulated claims bargaining gives players perverse incentives to demand everything.
• Refusing to prepare for limited war can be a deliberate deterrent strategy
• Restrained competition everywhere favors discrete qualitative over graded limits
• Salvo attacks require reconnaissance to know which targets are already destroyed
• Secret ballot strips voters of power to prove their votes
• Secret ballots protect voters by destroying evidence of their vote
• Security against false alarm and against surprise have opposite externalities
• Selecting visibly irrational agents can make threats credible
• Sequential-game solutions collapse into iterative elimination of dominated strategies.
• Side payments convert pure games into mixed bargaining games
• Simple discrete salient limits enable coordinated expectations.
• Slowly escalating punishments prove more credible than sudden fatal threats
• Small first-strike temptations can compound into attack through interacting expectations alone
• Small proportionate sanctions earn more credibility than dramatic all-or-nothing ones
• Small-probability threats are superior to large certain threats in compellence
• Some preferences are inherently unknowable even when their owner is desperate to reveal them
• Splitting a one-shot deal into installments creates artificial continuity
• Stability of mutual-deterrence equilibria depends on the slopes of each side’s response functions
• States want the enemy to verify our own peaceful intentions.
• Strategic moves work by constraining one’s own behavior to shape an opponent’s expectations.
• Strategic principles in mixed-motive games cannot be derived purely analytically
• Strategic wins always operate through constraining the opponent’s choice.
• Subjective reservation values cannot be credibly communicated even when honest
• Suggestive background stories can shift which equilibria players settle on
• Superior resources and skill can actually disadvantage a bargainer
• Tacit agreements need qualitatively distinct terms that may discriminate against one party
• Tacit and explicit bargaining form one continuum dependent on coordinated expectations
• Tacit bargaining forces one party to accept an unfair focal point
• Tacit bargaining over distribution can be conducted only through play itself.
• Tacit bargaining requires coordinating on equivalent solutions, not just equal-valued ones
• Tacit coordination is guess-the-guesser recursion, not prediction of average behavior.
• Tacit coordination on an outcome requires a distinguishing selection device
• Tacit coordination supplies the empirical premise Nash solution concepts require
• Tacit coordination works through mutually recognized focal points
• Tacit limits require qualitative boundaries amenable to complete abstention
• Tacit-bargaining theory guides where to seek terms, not their probability
• Technical weapons characteristics no longer distinguish nuclear from conventional arms
• The bargaining solution’s appeal derives from its focal quality, not its formal derivation
• The esprit de corps of social units persists via self-reinforcing convergent expectations.
• The first extant proposal dominates when coordination is hard in tacit games.
• The power to send but not receive gives a bargainer an unanswerable advantage
• Third-party enforcement can rely on conditional self-identification
• Threat credibility depends on the other party recognizing the retaliation formula
• Threat of force differs from application of force
• Threatening a small risk beats threatening certainty when the enemy’s controllability is uncertain.
• Threats and promises are meaningless in zero-sum games
• Threats fail against irrational adversaries and trapped opponents
• Threats gain credibility by decomposing into a series of smaller demonstrated steps
• Threats only matter when players can irreversibly change the game
• Threats require commitment, not comparative damage to both parties
• Threats, promises, and commitments unify as payoff-restricting reaction functions
• Time preferences beyond uniform discount rates let players alter the game by delaying
• Total prohibition beats partial limits as a tacit agreement focus
• Treaty most-favored clauses can strengthen a weaker ally’s bargaining position.
• Unforeseeable escalation risk alone can deter aggression even among cautious rational actors.
• Unilateral, non-reciprocal communication can coordinate both parties’ expectations
• Unstable focal points serve as signals rather than as outcomes
• Unwritten law rests on fear that breaking one rule collapses all
• Value systems are generally incommunicable except in special cases
• Value systems of different kinds of people resist introspective estimation
• Verification systems serve either truth-detection or credible self-disclosure
• Visible exogenous distrust parameters still leave amplifying interaction to explain
• Warning is only as valuable as the pre-planned response it triggers
• Warning-time gains must be exploited via pre-planned negotiation or wasted
• Warnings differ from threats because warnings convey true incentives without commitment
• Weapons designed for first strike destabilize deterrence by inviting preemption.
• Willful inattention lets enforcers avoid incurring an obligation to punish.
• Without legal enforcement or reputation audiences, some bargains become impossible
• Worsening one’s own outcomes can be strategically advantageous in mixed-motive games.