Classical Realism: the root of the tree

Thucydides: The Melian Dialogue

c. 416 BCE · extract from The History of the Peloponnesian War

In 416 BCE, during the war between Athens and Sparta, Athens demanded that neutral Melos, a small island colonized from Sparta, surrender and join its empire. The Athenians argue that power is what matters, not abstract justice: “the strong do what they can, the weak suffer what they must,” and letting Melos stay neutral would invite other subjects to rebel. The Melians appeal to justice and fairness even between unequal states, to the hope of Spartan rescue, and to divine favor. The Athenian case is blunt, pragmatic, and ruthless; the Melian case is moral and honorable but unrealistic given their weakness and isolation. Melos refuses, is besieged, and is destroyed: the men killed, the women and children enslaved. The ending is shocking, and it shows starkly how power politics can override morality in war. But I don’t read Thucydides as endorsing Athens. The larger history reminds us that Athens eventually lost the Peloponnesian War; many read the dialogue as a warning about arrogance and over-expansion. So it is not just a statement of power politics but also a tragedy: Athens followed the logic of power, and that very logic contributed to its downfall.

He is called a realist because he shows international politics as driven by power, fear, and interest rather than morality, Athens says openly that justice only holds between equals. Reading it today introduces the foundational problem of power versus morality, and it stays in conversation with the realism-versus-idealism/constructivism debate and with current issues: neutrality, imperialism, humanitarian law, and the limits of moral argument in global politics.

Carr (1939): The Twenty Years’ Crisis

Ch. 1 & 5 · the realist critique of utopianism

For Carr, political science mixes purposes (norms, what ought to be) with analysis (facts, power, what is); it is deeply purposive, created not just to understand the world but to change it. The field grew from the popular demand after WWI to prevent another catastrophe, and the modern “science” was born when post-war utopian hopes collided with the hard realities of the 1930s. I partly agree, though his origin story leans Anglo-European and underplays earlier, non-Western statecraft. Utopianism, in his terms, puts wish before analysis, believes in a universal harmony of interests, and derives practice directly from abstract ethical theory, trusting reason and morality to tame power. He would file Plato, the free-trade economists, early socialists, and above all Wilson and the League under this heading. Carr’s realism centers on power and interest: a hard analysis of facts over wishes, the relativity of all thought to the thinker’s circumstances, and the claim that practice creates theory (theories are often invented to justify interests). He traces it to Machiavelli, Hobbes, and Marx. The “harmony of interests” is the claim that what benefits a leading status-quo power benefits all, free trade, existing rules, peace under current arrangements, which Carr reads as ideology dressing particular interests as the common good. “Internationalism,” likewise, is a slogan dominant powers use to legitimize control (French internationalism under Louis XIV, British under the 19th-century empire, the League as an Anglo-French tool). Both critiques are compelling as a warning, always ask who gains, but incomplete: institutions can create real positive-sum gains and accumulate authority that gradually binds power.

His specific conclusions are tied to a unique context, the collapse of a particular liberal order across the “twenty years” crisis,’ so I’d be cautious generalizing them. But his method travels: be skeptical of claims to universal morality and examine the power interests behind them. His real meta-claim, that IR thought must join utopia (normative purpose) with realism (analysis of power), is the durable contribution. In my view Carr is mainly about dismantling utopianism, whereas Morgenthau goes on to build a positive grand theory.

Morgenthau (1948): Politics Among Nations

Ch. 1 · the six principles of political realism

Politics is governed by objective laws rooted in an unchanging human nature; the master concept is interest defined as power, an objective category that makes politics an autonomous sphere; that interest is universal but its content shifts across time and culture; political ethics differ from universal morality, with prudence the supreme virtue; no nation should equate its aspirations with God’s will. Both Carr and Morgenthau see politics as a struggle for power and distrust moralizing rhetoric, but Morgenthau is more systematic and roots his theory in a permanent human nature, where Carr is more historical and sociological. Carr mainly critiques utopianism; Morgenthau builds a positive theory of how states do and should behave. Human nature gives theory a foundation that feels objective and timeless, a simple explanation for conflict (Morgenthau’s animus dominandi; Schmitt’s friend/enemy), and a normative basis (Hobbes → strong sovereign; Locke → limited government; liberal idealists → trade and law secure peace). Its merits are explanatory parsimony, a warning against naïve utopianism, and the autonomy of politics. But it has real costs: an essentialist trap (“human nature” is defined selectively to fit the conclusion, why privilege aggression over empathy?), no room for history or culture, a conservative bias (if war is human nature, peace looks impossible), and unfalsifiability (cooperation is always dismissed as temporary).

Morgenthau means we can understand and predict a state by assuming its primary goal is to acquire, keep, and display power relative to others, a better guide than stated motives or ideology. I partly agree: it is a useful default for great-power security, but leaders also chase domestic survival, ideology, status, and personal goals. The theory’s weakness is definitional: “power” runs from physical violence to subtle psychological ties, so it is hard to measure and risks explaining everything and therefore nothing; it is better at explaining the past than predicting the future, and it blurs positive claims with norms (prudence is a virtue; policy ought to be rational). Reducing realists and idealists to pessimists and optimists is too simple. The real difference is methodological: how each frames international events. Realism emphasizes power, interests, and the constraints of anarchy; idealism emphasizes law, norms, and institutions. They are not mutually exclusive but complementary lenses, Jervis’s security dilemma sits exactly on the overlap, grounded in realist anarchy yet pointing to reassurance and cooperation. If every state pursued only narrow self-interest, the UN, EU, and WTO would never have become so influential. Pragmatically, I use theories for their explanatory value rather than forcing them into a binary. This is also why the debate still matters: realism supplies the field’s null hypothesis, power and self-interest as the default drivers, against which liberalism and constructivism must test themselves, and it maps directly onto real choices between engagement and containment, interventionism and isolationism.

Jervis (1976): Perception and Misperception

Ch. 3 · deterrence vs. spiral, and the security dilemma

Jervis writes in 1976, mid-Cold War détente, in the shadow of Vietnam, the nuclear arms race, and the near-miss of the Cuban Missile Crisis, which shifts the focus from the objective struggle for power toward the subjective psychology of leaders. The deterrence model assumes the adversary is aggressive and expansionist, sees the world as interconnected so that any concession signals weakness, and prescribes firmness and credible threats (the U.S. reading of Cuba). The spiral model assumes the adversary is fundamentally insecure, sees a world of security dilemmas where defensive build-ups look offensive, and prescribes reassurance to break the cycle (the pre-WWI Anglo-German naval race). Both can be right under different conditions, but I find the spiral model the more common danger: most states cannot reliably tell defense from offense in each other’s behavior, and most prioritize their own security. Jervis grounds the spiral’s bad outcomes not in human nature but in a “Hobbesian state of nature”, the structure of anarchy. With no authority to protect them, states must self-help, and self-help looks threatening to others, triggering the dilemma; the problem is in the situation, not the soul of the actor. This is the decisive break from Morgenthau, for whom the drive for power is an innate animus dominandi. Jervis is therefore more hopeful: change the structure or manage perceptions, and you can mitigate conflict.

The security dilemma is a permanent background condition of anarchy, but its intensity varies. It worsens when offense has the advantage and when offensive and defensive weapons cannot be distinguished, so states must assume the worst. It eases when defense has the advantage (nuclear second-strike makes homeland defense near-absolute), when diplomacy and transparency let states signal defensive intent, and when institutions reduce uncertainty. The 1972 ABM Treaty is a good example: by banning nationwide missile defenses, both sides preserved mutual vulnerability (MAD) and removed the incentive to race for defenses that would have triggered a new offensive build-up.

Johnson & Tierney (2019): “Bad World”

International Security · the negativity bias in international politics

The core claim is that “bad is stronger than good”: a deep negativity bias makes threats, losses, and failures loom larger than gains and successes, which helps explain threat inflation, the security dilemma, escalation, missed cooperation, and why organizations remember failure more than success. It also explains how overconfidence coexists with fear: we overweight negative information about others but positive information about ourselves. As a psychological principle it is well supported (hundreds of studies, neuroscience, cross-cultural evidence, an evolutionary rationale). As an explanation of international politics it is weaker: the authors lean on correlation more than mechanism, and the jump from individual psychology to state behavior involves aggregation problems they don’t fully resolve. A “rationalist” argument has actors weigh costs and benefits symmetrically and respond to incentives, treating apparent bias as optimal strategy. Against that baseline: (1) Threat sensitivity, actors overweight hostile cues and infer worst-case intentions, fueling threat inflation and the security dilemma (Soviet force cuts in the 1950s–60s did not soften Washington’s read of Moscow); a rationalist calls skepticism sensible under uncertainty, but the bias account adds that even favorable signals get underweighted. (2) Loss aversion, under losses, actors gamble rather than accept sure pain, driving escalation (the 2007 Iraq “surge” rather than accepting loss). (3) Failure salience, organizations overlearn from recent failures (Somalia 1993 helped deter action in Rwanda 1994). Mitigation matters: bias is worst in low-information settings, in judgments about others, and under recency; decision structures and routines, free-flowing information, and institutional checks help most. Even robust democracies can fall into collective bias, as with post-9/11 threat inflation. The WWI case study (German encirclement fears, refusal to accept defeat, post-Versailles fixation on failure) usefully shows the three biases compounding over time, though it feels rushed and doesn’t rule out strategic calculation.