Causes of War
This week works across more than one level. Levy (1998) maps the field across systemic, societal, and individual images and argues we should drop the “paradigm wars” for testable, mechanism-first, mid-range theory. Fearon supplies the rationalist core: war is a bargaining failure traced to information and commitment problems. Levy (1989) presses the domestic image with the diversionary theory. Lake (2010) takes bargaining theory to the 2003 Iraq War and finds the decisive failures were self-delusion, not opponent deception, a call for a behavioral theory of war. Goddard adds that indivisibility is constructed, not given. My own claim is that real wars are layered: a structural motive (often a commitment problem) erupting through cognitive failure and domestic mobilization.
Reading list
- Levy. 1998. “The Causes of War and the Conditions of Peace.” Annual Review of Political Science.
- Levy. 1989. “The Diversionary Theory of War: A Critique.” In Handbook of War Studies.
- Fearon. 1995. “Rationalist Explanations for War.” International Organization.
- Goddard. 2006. “Uncommon Ground: Indivisible Territory and the Politics of Legitimacy.” International Organization.
- Lake. 2010. “Two Cheers for Bargaining Theory: Assessing Rationalist Explanations of the Iraq War.” International Security.
Levy (1998): The Causes of War and the Conditions of Peace
Annual Review of Political Science · a levels-of-analysis review essay
Scholars are preoccupied with the causes of war and peace because (i) real-world shocks constantly raise the stakes of getting the explanation right; (ii) war is rare at the dyadic level and peace is a “non-event,” so we need strong concepts and methods to study both; and (iii) settling this adjudicates huge theoretical claims, realist, liberal, psychological, about how world politics works. “War” is large-scale organized violence between political units, conventionally measured by the Correlates of War threshold of at least 1,000 battle-related fatalities; “peace” is analytically distinct from justice and usually defined simply as the absence of war. Types: interstate, civil, and imperial/colonial wars, with attention shifting after the Cold War toward low-intensity and identity wars. I’d also flag a tradition the “paradigm wars” neglect, the Marxist-Leninist one (Lenin, Wallerstein, dependency theory, Gramscian IPE): war and peace follow the contradictions of capitalist development, the world economy creates core–periphery hierarchies, and power operates through economic and ideological domination, not just military force. Useful, because it forces multi-level thinking and guards against attributing a dyadic outcome to individual traits alone, and since most wars involve the mutual, interactive decisions of two or more adversaries, any explanation logically requires dyadic- or systemic-level variables. It is less useful if treated as a paradigm silo, which distracts from testable within-paradigm variation. Systemic level: realism (security dilemma, conflict spirals, iterated Prisoner’s Dilemma, Waltz), power-transition / hegemonic theory, and liberal interdependence; Levy is skeptical that bipolarity is more peaceful (the evidence is thin and Cold War confounds abound), and he suggests balance-of-power and hegemonic theories may be reconcilable across nested systems, the compelling insight being that regional wars escalate to global wars only under a deconcentration of naval power and wealth. Societal level: diversionary theory (Rummel), democratic constraints (unsuccessful war → leaders thrown out), and Snyder’s logrolled-coalition overexpansion; debates over interdependence and war are mixed and endogenous, and Levy & Barbieri’s finding that war doesn’t sharply depress trade demands domestic mechanisms, the democratic peace came “as close as anything to an empirical law,” pushing the field toward societal variables. Individual level: operational code, misperception, bounded rationality (George, Jervis, Lebow, Holsti), and prospect theory; the worry is that reference points are hard to identify and risk observational equivalence with expected-value models, but the payoff is clear, testable predictions, risk acceptance in the domain of losses, deterrence easier than compellence, cooperation easier over gains than losses.
A good review maps the field with a clear organizing framework (he uses levels of analysis so readers aren’t lost), evaluates mechanisms rather than mere correlations, checks the fit between research design and theory (he repeatedly flags misspecification and endogeneity, e.g., in trade–conflict tests), and points to tractable research gaps. Even a review makes arguments, and Levy’s are: (1) shift from paradigms to testable mid-range theory, he criticizes the “paradigm wars” for distracting from within-paradigm variation; (2) levels clarity and aggregation caution, use levels carefully and don’t extrapolate across them; (3) re-assess systemic staples, he doubts polarity explains war and peace; (4) push mechanism and methods, model the war↔trade simultaneity and bring in domestic variables. I’d add: the deepening of rationalist explanations (Fearon 1995, Powell, by the mid-2000s a dozen-plus formal models of war); the civil-war and ethnic-conflict revolution (the Ethnic Power Relations dataset, 2009, diverse societies aren’t inherently conflict-prone; the danger is exclusionary power configurations); territorial-peace theory challenging the democratic peace (peace may create democracy, not the reverse, and early democratization can raise conflict risk); the changed nature of war (interstate conflicts in the early 2020s already matching the entire 2000s, and non-state armed violence overtaking interstate war as the dominant form); and positive-peace concepts that treat peace as a measurable process, not just the absence of war. Methodologically I’d add causal inference (DID, IV, RDD, synthetic control), text-as-data, and geospatial event data, plus cyber, drones, gray-zone coercion, and sanctions. The “return to the first image” refers to Waltz’s three images (individual, state, system); after decades of third-image dominance, the field is re-attending to leaders, driven by strongman politics (Putin, Trump, Xi), by structural theory’s failure to predict Russia’s 2022 invasion, and by better leader-assessment tools. But it isn’t a simple return: personality matters most in crises and under scarce or ambiguous information, yet its relevance still depends on systemic constraints and domestic structure. My own bottom line is an integrated multi-level view. Leaders’ beliefs and biases matter at the first level; democratic transitions and horizontal (ethnic) inequalities matter at the second; anarchy and commitment problems matter at the third. The Russia-Ukraine war runs through all three.
Levy (1989): The Diversionary Theory of War: A Critique
Handbook of War Studies · the scapegoat hypothesis and its conditions
The idea is that political elites embark on adventurous foreign policies or even war to distract popular attention from internal social and economic problems and consolidate their own domestic support, seeing war as a “necessary or convenient means to establish, maintain, or expand the power of a government, party, or class.” Three logics run through it: “wag the dog” (distraction), “rally round the flag” (cohesion), and gambling. My prior was the simple version: politicians in trouble pick fights abroad to rally support, which Levy validates but complicates: not every externalization of conflict is scapegoating, and the conditions under which diversion “works” are narrow. Two examples I find apt: France and Russia before WWI in the July Crisis of 1914 (Poincaré promoting nationalism to wipe away the shame of 1870 and shore up his position; the Tsar using Pan-Slavism after the failed 1905 revolution, so abandoning Serbia would look like betraying the Slavic cause; both were far readier to risk war than Germany expected); and the Yuan dynasty’s invasions of Japan (1274, 1281), where an alien Mongol regime with fragile legitimacy among both Mongol nobles and Chinese subjects used external campaigns to project unity. Simmel and Coser argue external conflict can raise internal cohesion, but only under conditions: the group must already be a “going concern” with minimum cohesion, the threat must be perceived as facing the whole group, and members must believe collective action can address it. So there is no universal “rally” effect, escalation won’t consolidate a fragmented society, a threat seen as partisan, or a government that lacks credibility. On conditions for diversion: regime type matters, in industrial democracies, economic downturns correlate with greater involvement in disputes (electoral punishment → scapegoat temptation), while in non-democracies involvement rises in expansions, and importantly, these links fade once you shift the outcome from “involvement” to “escalation to war.” Power position matters too, major and especially leading powers face fewer external constraints, so they can exploit diversion more freely. And some hardline behavior is not diversion at all but the redistribution of rents to cement coalitions. My synthesis: diversion is likeliest when (i) leaders face acute but manageable domestic risk (a bad economy, scandals) and expect a limited move to generate a rally; (ii) the target is weak enough to keep costs contained yet salient enough to mobilize identity; and (iii) the opponent’s reaction is expected to be mild, or the leader misjudges it. A declining state with serious domestic problems may combine preventive and scapegoat motives into “gambling for resurrection.”
The diversionary theory is hard to evaluate because causation runs both ways and timing is tangled: internal conflict may drive external conflict (diversion), but external conflict can also raise or dampen internal conflict (rally vs. backlash), so a same-period correlation of “domestic unrest” with “international conflict” misses lags and feedback. Levy’s critiques of the quantitative literature are, I think, fair: it often counts proxy events (protests, sanctions, diplomatic notes) instead of the theorized outcome (war or militarized behavior chosen under domestic pressure); it ignores directionality and reciprocity; it picks essentially arbitrary time lags with no theoretical basis; it uses static, cross-sectional, correlational designs for a claim that is temporal, dynamic, and causal, “flawed from the start”; and its samples are often too narrow (e.g., mid-1950s–1960s) to generalize. The best approach models dynamics and simultaneity (simultaneous equations, panel VAR, event history), disaggregates outcomes by intensity (disputes vs. war), conditions on context (regime type, great-power status, power trends), and uses structured, focused case comparisons with process-tracing of cabinet records, speeches, and budget flows to pin down motive, embedding the opponent’s reaction with a Fearon-style bargaining model, as Levy stresses. I’d combine a two-stage selection-and-escalation model with dynamic panels and process tracing. Even restricting attention to internal conflict leading to external conflict, several distinct mechanisms could be at work: (1) diversion/scapegoating, the classic; (2) opportunistic intervention by others, internal strife in State A creates weakness that State B exploits (Iraq attacking Iran in 1980); (3) outcome-shaping interventions, strong states intervene to steer A’s domestic struggle (the US in the Dominican Republic, 1965; the USSR in Czechoslovakia, 1968, and Afghanistan, 1980); (4) a combined process, A’s strife tempts B to attack, and the attack then hands A’s leaders a real external threat to exploit (revolutionary regimes: France 1792, Russia 1918, Iran 1980); and (5) Blainey’s finding that in many 1815–1939 wars the strife-torn nation did not initiate, strife weakened it and invited attack by upsetting deterrence. I find them convincing, empirically and logically: they compete with the diversionary story and warn against reading every internal–external link as scapegoating, which is exactly why Levy insists on mechanism-specific tests.
Fearon (1995): Rationalist Explanations for War
International Organization · war as a bargaining failure
“Unitary, rational actors” means states are modeled as single decision-makers maximizing expected utility given their beliefs and constraints, “billiard balls,” with no internal coalition bargaining or principal–agent frictions inside the state. Fearon uses this to build an ideal-case benchmark, so we can pinpoint the minimal, logically coherent reasons war can occur even when leaders are instrumentally rational; understanding what rationality really implies may actually raise our estimate of how much irrational and second-image factors matter. A rationalist approach keeps that benchmark and shows war can happen even when both sides weigh costs and benefits; a non-rationalist approach leans on biases, pathologies, or leaders who reap benefits while pushing costs onto others. War is “puzzling” because it is ex post inefficient, it destroys value, so rational states should be able to find a negotiated settlement ex ante that both prefer to fighting (the bargaining range). The puzzle is what prevents them from locating and agreeing on a bargain inside that range. Note that some credibility tools, like domestic audience costs, actually sit outside the pure unitary frame, because they presume leaders answer to a domestic audience. (1) Information problems: leaders hold private information about capabilities or resolve and have incentives to misrepresent it, to exaggerate so as to deter future challenges or win concessions, or to conceal so as to protect a first-strike option or avoid looking like the aggressor. Cheap talk isn’t credible (if A conditions its demand on B’s statements, B says whatever lowers the demand, so A learns nothing); and even costly signals can raise war risk, because to be informative a signal must be costly in a way a weaker type would avoid, mobilizations, audience costs, limited deployments, which create real escalation danger. Formally, A’s optimal demand carries a positive ex ante probability of war. Example: the July Crisis of 1914. It’s theoretically tight; Fearon notes leaders rarely update on capabilities during crises, but updating on resolve is very common. (2) Commitment problems: under anarchy, with no third party to enforce agreements, some dynamics make credible promises impossible, anticipated power shifts, strategic territory affecting future bargaining, first-strike advantages. A declining state may prefer to fight now rather than accept a worse peace later that the rising state can’t credibly promise not to demand, the logic of preventive war. Example: Fearon’s own reading of the 1990–91 Gulf crisis as preventing an oil hegemon’s future leverage, fear of the future peace, not of invasion. Strong on internal logic and tractable around exogenous timing. (3) Issue indivisibility: some issues seem to admit only a finite set of outcomes (thrones, sacred sites, Jerusalem), so none may lie in the bargaining range. But Fearon is skeptical.
Because most real disputes are complex and multidimensional, so side-payments, issue linkage, alternation, or randomization usually exist in principle; effective indivisibility therefore tends to be endogenous to politics (norms, nationalism, domestic audiences) rather than inherent in the issue. When fully developed, the diverse rationalist explanations collapse into the two deep mechanisms, private information with incentives to misrepresent, or commitment problems. I agree as a general theory of war. Indivisibility bites in specific cases (sacred sovereignty, identity-laden symbols), but when you probe “why not compensate or link?” you usually uncover either incentives to misrepresent or commitment problems. The sharpest version is elite manufacture: leaders cultivate a mass belief that a symbolic issue is non-negotiable, divisible to the elites, “indivisible” to the public. That “effective indivisibility” reduces to (a) strategic misrepresentation (claiming the issue “can never be bargained” to raise leverage, while opponents can’t tell genuine constraint from posturing) and (b) a domestic-audience commitment problem (even if elites would secretly compromise, they can’t guarantee the public won’t overturn the deal). Later work runs exactly this way: Fearon’s own 1994 audience-costs model institutionalizes “effective indivisibility” by tying leaders’ hands; Snyder (Myths of Empire), Kaufmann, and Brubaker & Laitin show elites inflaming nationalism and identity to make divisible issues seem sacred; and constructivist IPE (Abdelal et al.) treats identity narratives as political capital whose rigidity determines perceived divisibility. Fearon brackets the domestic level deliberately, not denying it, but insisting we first fix the rational baseline before bringing in non-rationality.
Goddard (2006): Uncommon Ground: Indivisible Territory and the Politics of Legitimacy
International Organization · indivisibility as a constructed, legitimation-driven phenomenon
She defines an issue as indivisible if actors represent it, territory, say, in ways that eliminate any possible division and thus reduce the bargaining range to zero: technically the values become discontinuous, and informally any division one side accepts the other rejects. Importantly, this departs from physical indivisibility, territory is physically divisible yet can still be treated as indivisible in practice. Her cases are Jerusalem, Ireland, Kosovo, and Kashmir. Indivisibility is a constructed social fact: neither an objective, inherent property of territory nor something subjective and reducible to individual consciousness, what matters is whether all actors reject mechanisms of division. It is contingent, a product of human action, yet “not easily malleable,” because once legitimation locks actors into a single coalition, structural constraints and audience costs make change costly and it hardens over time into a hegemonic discourse. I agree: treating indivisibility as endogenous to interaction explains temporal variation, why some periods look “all-or-nothing”, though the measurement burden is real, since the mechanism lives in public justifications and coalition structures. Whether territory appears indivisible depends on how actors legitimate their claims during bargaining; following network theory, legitimation strategies transform networks of coalitions, producing “switching effects” that build or sever ties and gradually lock actors into positions where they can no longer recognize any other claim as legitimate, eliminating the bargaining range. A legitimation only works if its public audience can hear and accept it, and discourse only resonates within culturally consonant networks. (1) Coalition mobilization: when a legitimation resonates, and ties are dense and the actor central, it activates and strengthens ties to a coalition, increasing dependence and reducing bargaining flexibility. (2) Polarization: when legitimations are culturally dissonant, with sparse ties and peripheral pulls, arguments sever ties, pushing moderates to the periphery and locking actors into a single coalition. (3) Yoking (actor creation): with an identity bridge in a bridging position plus culturally inventive rhetoric that recombines symbols, entirely new actors or identities can be created during bargaining. My reaction: mobilization is theoretically tight and well grounded in network logic (density, centrality); polarization nicely explains how minority extremists shape outcomes when embedded with ruling parties; yoking is the boldest, it explains identity and actor creation (e.g., “Ulster” as a coherent actor), but it is the hardest to measure cleanly. Mobilization and polarization together feel most compelling, mapping directly onto observable tie changes.
Ulster is a hard, historically important case widely thought “naturally indivisible,” so showing it was constructed is a strong plausibility test. The turning point is the 1886 Home Rule debate, when Conservatives “played the Orange Card,” reframing the dispute from imperial integrity to the defense of Protestant rights in Ulster. Mobilization: that frame activated and strengthened ties between Conservatives and Protestant “Ascendancy” networks, building them as guardians of Protestantism. Polarization: the Orange Card was dissonant with British Liberals and Irish moderates, letting Conservatives brand them as illiberally “coercing” Ulster and sever ties with moderates, pushing the dispute into incompatible claims. Yoking: “Ulster” did not always exist as a coherent actor, the British deployment of “Ulster Protestantism” yoked the identity together. Lock-in: as ties concentrated on Ulster, Conservatives became structurally dependent, so recognizing any Irish national claim would undermine their own legitimacy, collapsing the bargaining range. Strengths: mechanism-heavy process tracing and a genuinely hard case. Weaknesses: legitimation and “resonance” are coded qualitatively from speeches and the press, so clearer coding rules or cross-case designs would help, and the article concedes legitimation isn’t the only cause (institutional reforms, security shocks, and electoral incentives also matter). I think the mechanisms travel to status, sovereignty, and identity-linked disputes where public legitimation and tie structures are central, Milosevic’s Kosovo rhetoric (mobilization) locking him into extreme Serbian nationalists, small extremist coalitions binding ruling parties (polarization), and Bismarck recombining nationalist and conservative rhetoric to bind a new nationhood (yoking). A useful modern counter-example: in 1992 Japanese and Chinese leaders kept Emperor Akihito’s visit from becoming a zero-sum struggle over war responsibility by using different, deliberately vague messages for each audience, making “history and apology” divisible rather than sacred, much as the fifteenth-century Ashikaga shogunate accepted the title “King of Japan” from Ming China while claiming full autonomy at home. Both show actors using bridging, depolarizing discourse to avoid constructing indivisibility. To generalize, I’d want multi-case comparison (Jerusalem/Kosovo/Kashmir) with standardized discourse-network coding.
Lake (2010): Two Cheers for Bargaining Theory
International Security · assessing rationalist explanations of the 2003 Iraq War
Lake opens by noting that, despite its importance, IR scholars hadn’t given the Iraq War sustained causal analysis, most accounts say how, not why, so he assesses the leading academic theory, the rationalist/bargaining approach. He reads the deeper issue as regional dominance in the Persian Gulf, with WMD the proximate trigger, and argues bargains were available ex ante (containment, forcible inspections, internal regime change were all feasible at lower cost). A preventive war follows from the commitment logic: if power will shift, the side expecting to be weaker later has an incentive to fight now to avoid a worse deal. Iraq looks preventive because the US feared a future WMD-armed Iraq would raise US costs and reduce US leverage, compounded by uncertainty over Saddam’s future intentions, so it chose to fight before the shift. Lake calls this necessary but insufficient: the actual bargaining failures were not those the theory predicts, and four standard assumptions need revision, unitary actor, two-player modeling, war ending at settlement with no postwar costs, and strong rationality. I agree the preventive story captures why early action could appeal, but it’s best read as a call to extend rationalist theory, not refute it. Bargaining theory expects private information with incentives to misrepresent to be central, actors hide their costs or capabilities, bluff resolve, and if learning fails, war occurs. Contrary to expectation, Lake finds there was relatively little private information about comparative capabilities or the likely battlefield outcome: US victory was near-certain and war plans were widely reported. Two information problems did matter, but in unexpected ways: Saddam misread US resolve, believing the US was a “paper tiger” that would not go to Baghdad (no Iraqi leaders believed coalition forces would reach the capital); and the Bush administration misread its own costs, treating postwar governance costs as small to nil, the single largest US failure in the bargaining. So the “information problem” here is not classic hidden-type bluffing by the opponent; it is misbelief about oneself plus misreading the other side’s resolve, both behavioral departures from Bayesian updating. I find that persuasive: misrepresentation by the other side was far less of a problem than self-delusion.
Domestic politics mattered because special interests, oil firms, the defense industry, and (citing Mearsheimer & Walt) the debate over Israel’s role, plus partisan politics could lower the perceived cost of war and make peaceful bargains harder. But Lake argues special interests could cause war by themselves only under unlikely conditions: that they fully control decisions, can bear extremely high war costs, and can override opposition in the public and bureaucracy. None held in 2003, the war was not fought for Exxon or Halliburton; these interests made a peaceful resolution more difficult but did not drive it. They are an amplifier, not the engine. “Behavioral” or “irrational” here doesn’t mean random action; it means systematic deviation from rational updating and full-cost calculation, ignoring foreseeable postwar governance costs, clinging to priors about adversary resolve. Since the decisive failures were self-delusion rather than opponent deception, Lake calls for a behavioral theory of war. My view tracks his hierarchy: domestic politics amplified hawkish choices and filtered information, but the decisive lever was the commitment problem as interpreted through misbeliefs about costs and postwar governance. Lake concludes the Iraq War can’t be fully explained by standard two-player bargaining with classic information/commitment terms; we must amend theory in four ways, integrate postwar governance costs ex ante, relax the unitary-actor assumption, move from two-player to multi-audience signaling, and incorporate behavioral decision biases. Are his critiques of Fearon fair? Partly. Fearon deliberately isolated rationalist mechanisms in a two-state setting and didn’t deny domestic politics or psychology, only bracketed them, so Lake’s points are best seen as extensions, not refutations; he’s strongest on postwar governance costs and behavioral updating, and least fair where he implies the originals claimed to be complete causal accounts rather than baseline models. What he undercounts, I think, is how much core preventive logic still pushed early action once leaders believed a future WMD Iraq would sharply worsen the bargain, plus back-channel diplomacy and regional third parties (Iran, regime survival) that a fuller n-player model could capture. The deeper payoff is that “rationality” is not one thing. Fearon and Powell stress instrumental rationality (optimal means to given ends), by that logic Iraq shouldn’t have happened, so war signals a failed mechanism (information, commitment). Lake offers procedural rationality, the war came from self-delusion and information distortion, so rationality itself failed. Jervis frames war as the “rational consequence of mistaken beliefs”, decision-makers stayed logically consistent inside their own belief systems (cognitive rationality). Constructivists like Campbell and Jackson go further, treating rationality itself as ideology, a self-legitimizing power discourse. Iraq shows rationality becoming ideologized: a post-9/11 “security vs. civilizational conflict” frame cast Saddam as anti-Western tyranny, the policy process bent all information to a preset conclusion, and “spreading democracy” and “pre-emption” became legitimizing rhetoric, rationality shifting from logical calibration to meaning production. So the concept remains useful as a modeling baseline (the counterfactual “if rational, no war” lets us ask what undermined rationality), but also as an entry point to how power and belief organize decisions. The US in Iraq was neither simply rational nor irrational but engaged in “rationalized action” bounded by belief and discourse, This helps explain why the causes of war are so hard to pin down: even a clean structural incentive runs through what leaders believe and how institutions filter information. Lake’s lesson is amended rationalism: same logic, richer assumptions.
Application · My extended analysis, three modern conflicts, the multi-party view of Israel–Palestine, and the limits of the bargaining dichotomy
Drawing on my own reports applying the bargaining / cognition / domestic frameworks to the 2022 Russia-Ukraine, 2023 Gaza, and 2025 Iran-Israel wars; on a multi-party narrative analysis of the Israeli–Palestinian conflict (2023–2025); and on a comparative study of alliance/security strategy. Background sources include CRS R47754 and CFR, IISS, and EUISS briefings.
First, three modern conflicts, separated by mechanism. Applying the rationalist, psychological, and domestic frameworks one case at a time shows their causes are not monolithic. Russia’s 2022 invasion of Ukraine is best read as a commitment problem driving a preventive war: Moscow saw Ukraine’s irreversible integration with NATO as a structural power shift, and because Ukraine had written NATO membership into its constitution while the West refused to “shut NATO’s door,” no credible commitment to permanent neutrality was possible, so Russia chose to fight before the strategic loss locked in. Information matters less here (the preventive logic holds even with perfect information, given NATO’s zero-sum expansion), and domestic consolidation may have set the timing, but “stop NATO expansion” has been a consistent, openly stated demand since the 1990s. The 2025 Iran-Israel war is also a commitment problem: with Iran at 60% enrichment, ~400 kg of highly enriched uranium, IAEA monitoring curtailed, and the JCPOA dead, Israel could not trust any verifiable promise to halt weaponization, so it struck during a “strategic window” to stop Iran crossing the nuclear “point of no return.” The 2023 Gaza war is the odd one out: although Hamas’s motive was structural (a commitment problem about blocking Saudi–Israel normalization), the enabling cause was Israel’s collective misperception, the “mow the grass” assumption, over-reliance on the border wall and surveillance, and the belief that Hamas wanted only governance and economic gain. The lesson is that resolving Russia-Ukraine and Iran-Israel needs credible, verifiable long-term commitment mechanisms, whereas the Gaza dilemma needs leaders to overcome cognitive bias and re-read adversary intentions and capabilities.
Second, the multi-party view of Israel–Palestine (2023–2025): narratives, power, and a fragile peace. Beyond mechanisms, the war is also a fight over narrative. Western mainstream coverage has been widely criticized as systematically downplaying Palestinian suffering and rationalizing Israeli military action; some scholars call this “epistemic violence” that normalizes power asymmetry and strips Palestinians of agency (e.g., truce coverage centering Israeli hostages while the release of hundreds of detained Palestinians drew little attention), compounded by Israel’s control of embedded reporting. Responsibility attribution splits along great-power lines: the US backs Israel and its right of self-defense and pushes a US-led order via the “Trump 20-point plan,” while China and Russia frame the root cause as the unresolved Palestinian question, demand an immediate ceasefire, call Israel’s campaign “beyond self-defense,” and support a two-state solution, positioning themselves as champions of the Global South, distracting the West from Ukraine and the Indo-Pacific and protecting Belt and Road interests, and jointly vetoing a US ceasefire resolution as too “ambiguous.” The October 2025 ceasefire came not from diplomatic consensus but from coercive power politics and sheer exhaustion (over 67,000 Palestinian deaths): the turning point was Israel’s September 2025 strike on Hamas negotiators in Doha, which threatened Qatar, a key US ally hosting a US base, leaving US negotiators feeling “betrayed” and forcing Trump to pressure Netanyahu into the deal, making the agreement a strategic pause rather than a real peace. Post-war governance is the deepest uncertainty: the 20-point plan is vague on disarmament, largely ignores Palestinian aspirations, and no country has volunteered for its International Stabilization Force, while Arab allies live a “dual reality”, publicly condemning Israel yet privately deepening security cooperation under CENTCOM, driven by the shared Iran threat, so any stabilization force risks being seen as an American-Israeli proxy and inheriting a severe trust deficit. Unless occupation, self-determination, and the details of disarmament are resolved, the ceasefire is inherently fragile.
Third, what the bargaining dichotomy misses, strategy beyond “balance vs. bandwagon.” A comparative study of alliance and security strategy sharpens the same point from the angle of how states choose under threat. Vietnam’s response to China shows the classic “balance vs. bandwagon” dichotomy is too thin: facing a proximate threat in the South China Sea that fits Walt’s criteria (proximity, offensive capability), Vietnam neither pure-balances (its “Three No’s” rule out a formal US alliance) nor appeases, but adopts an omnidirectional hedge, economic bandwagoning-for-profit with China (in Schweller’s sense) plus military soft-balancing and partnership diversification with the US and others. This challenges Walt (threat perception must include the cost of resistance and economic dependency, not just military capability and intent), confirms Schweller in the economic domain, and shows Risse-Kappen’s ideological-mitigation mechanism failing among non-democracies, Vietnam and China share a Communist-led system yet remain rivals, so shared ideology does not create a “presumption of amity.” It supports Kang and Spektor on hierarchy and Global-South agency: small states hedge to “play the great powers off one another” and preserve strategic autonomy. My proposed revision is to add an economic-leverage/dependency variable to the balance-of-threat model and to treat hedging as an independent third option on the risk continuum between balancing and bandwagoning, a reminder that the same threat-perception and security-dilemma machinery driving the causes of war also drives the alignment choices that decide whether those wars stay contained.
← Back to the genealogy tree · ← Prev: Week 7, Regime Type & Conflict · Next: Week 9, Alliances →