In a recent interview, Hilary Clinton described her likely general-election opponent, Donald Trump, as “dangerous.” Seasoned observers often shrug off this kind of hyperbole, especially in an election season.

But ideas of political danger have changed throughout U.S. political history — and what politicians describe as “dangerous” tells us something about the stability of our political institutions.

A little historical background will show that the concept of political danger is being used today in a new way.

In the early republic, the danger was civil war

Internal divisions were the most pressing threat facing the new republic at the founding. The framers of the Constitution worried about the political danger of political parties (or “factions,” as they were sometimes called in the 18th century) — in no small part because they feared that parties could tear the union apart, dismantling the republic entirely. The federal government had limited resources and personnel scattered over a vast territory, making it, as historian John M. Murrin put it, a “midget institution in a giant land.”

Statesmen and political elites knew that, in a free society, they couldn’t stop parties from forming. But they feared them because party feeling could be hot-tempered and even violent. Parties presented a problem because governing authorities were ill-equipped to contain the threat that partisan rivalries posed to civil peace and the unity of the republic.

In fact, as the French Revolutionary and Napoleonic Wars raged in Europe from 1793 to 1815, U.S. statesmen realistically feared a proxy war between France and Britain on American soil, with the Democratic-Republicans, a Francophile party, at odds with the Federalists, an Anglophile party, over questions of war and peace. The French minister, Edmond Genêt, tried to mobilize Democratic-Republicans to launch hostilities against the British. John Adams later, in 1813, wrote to Thomas Jefferson:

You certainly never felt the Terrorism excited by Genêt, in 1793, when ten thousand People in the Streets of Philadelphia, day after day threatened to drag Washington out of his House, and effect a Revolution in the Government, or compell it to declare War in favour of the French Revolution and against England.

Parties were still a threat during the years before the Civil War. Politicians and power brokers were then trying to prevent the parties from setting off a conflagration between slaveholding and non-slaveholding states. In the wake of the Missouri Crisis over whether slavery should extend into the territories, political leaders reasonably feared that if parties divided the nation regionally, they would threaten the Union.

After the Civil War, a new idea of political danger emerged

As I’ve argued elsewhere, observers after the Civil War no longer feared that political parties would pull the Union apart – and parties came to be seen instead as organizations vital to a modern democracy.

After the First World War and the Bolshevik Revolution, political groups (rather than political parties) were defined as dangerous. Ethnic and labor groups were attacked for allegedly sympathizing with foreign powers and promoting collectivist — and anti-democratic — ideologies.

Now the danger is an unpredictable U.S. federal government

Observers who allege that Trump is a dangerous candidate use the concept of political danger in a new way. Many posit him as a threat to the rule of law because his behavior does not appear to be governed by a familiar set of rules.

Consider the way this concept is used in a recent Dana Milbank op-ed in The Washington Post. Milbank cautioned Democrats who minimize the potential threat of Trump’s candidacy. Sen. Ted Cruz, Milbank argued, would have been better because he follows a logic that is intelligible to others: He’s governed, first and foremost, by ambition. As Milbank wrote, “Cruz believes in Cruz” and is a “creature of democratic institutions.” That makes him predictable and amenable to rules and limits.

Milbank’s reasoning might be extended one step further. Not only are ambitious politicians more predictable, they are motivated (when it serves their interests) to build durable coalitions and to cultivate meaningful political relationships with allies.

Would a Trump presidency be a rogue presidency, unaligned and unpredictable?

Trump, by contrast, is his own brand, neither a leader of nor affiliated with an established, organized group. He has not been meaningfully bound to the Republican Party, the party he will apparently lead in the fall. He does not espouse a specific or consistent political message or persuasion. He seems unencumbered by the need to build coalitions and foster alliances — “commitments of interest and ideology,” to borrow an expression from political scientist Stephen Skowronek — that we’ve come to expect of presidential candidates.

Even the alleged demagogues of the past relied on political parties, allied with organized groups, and built relationships with politicians and power brokers. By contrast, a Trump presidency, at least as others describe what it would look like, would be a rogue presidency.

A rogue presidency would be particularly threatening in today’s economic and geopolitical climate. Whoever steps into the Oval Office inherits a vast array of policy commitments and relationships of mutual interest with domestic and foreign stakeholders — commitments that require steady management. U.S. presidents give signals and cues that communicate important information about how the federal government will deal with adversaries and allies and domestic and foreign markets.

Today we no longer worry about internal dangers to the structural integrity of the Union. So what are the political dangers we fear?

The United States is now a structurally vital part of a densely networked global economy rife with security challenges. To keep that economy turning, our political assumption is that U.S. presidents should be calculating individuals — and should follow a calculus that is intelligible to others.

By contrast, Trump argues that his unpredictable persona is his strength. It does serve him politically. He regularly departs from his script in ways that both display authenticity and attract news media attention.

But how would such a persona play on the international stage? The personality of a president (or of a presidential candidate) has taken on a new institutional significance. U.S. allies, partners and other global audiences are worrying about a historically new political danger: the danger of more uncertainty in an already fractious and unpredictable world.

Jeffrey S. Selinger is assistant professor of government and legal studies at Bowdoin College in Brunswick, Maine, and author of “Embracing Dissent: Political Violence and Party Development in the United States.”