To draw a bead on threats to democracy in the second Trump Administration, one can do worse than watch a video of an arrest made by ICE in the sleepy college town of Evanston, IL, home to Northwestern University, at the beginning of November 2025. The video shows a protestor being arrested after his car rear-ended an ICE vehicle. The ICE agents have the man restrained on the ground. They then start punching him in the head, repeatedly. Toggle over to YouTube, and scan the comments, and what’s striking are the number of people, presumably Americans, cheering on a violent government assault on a fellow citizen, an assault that could have easily ended in a mortal tragedy.

This snapshot of the new America might be invoked as an exhibit on the case for several diagnoses of democratic erosion. Usefully, the incident also helps us see how such erosion unfolds along different planes and that the interaction between those various threads can induce vicious circles of democratic unraveling. 

The Evanston incident might serve, to begin with, as exhibit A in the case for mapping a larger failure of legal constraints on federal power, and in particular federal violence. As Tom Ginsburg and I argued in a paper first posted in January 2017, effectual rule-of-law constraints on police, prosecutors, and state more generally are a sine-qua-non of democratic competition. Democracy where officials feel free to arrest or beat the opposition, or stuff ballots, is usually a charade. Earlier this year, I argued that this Administration was deliberately cultivating a version of what Ernst Fraenkel called “the dual state,” by which officials have a variety of tools for simply ignoring, or effectively switching off, the law. The difficulty Judge Sara Ellis has had eliciting compliance with her court order restraining ICE violence against citizens and noncitizens alike in Chicago is evidence that the observation, and Fraenkel’s model, has some staying power now. Of course, one might also point to prosecutions of the president’s perceived political foes or partisan misuse of regulatory power to punish other political opponents as further evidence of the same point.     

Alternatively, the Evanston incident and reactions to it showcase the deep affective polarization cutting Americans off from each other. Political violence has plagued public figures on both the left and the right, so one might imagine a mutual interest in a cooling of rhetoric, a search for the common ground evinced, for example, in Beth Macy’s powerful new book.  But, with dismaying predictably, it seems there are many Americans who orient their political identity around a hatred of the other party. Certainly, anyone who writes for the general public (or blogs, or tweets, I suppose) from a democratic, generally progressive perspective, is well aware of the depth and extent of violent vitriol on the right. Polling data released mere days after the Evanston incident finds a sizable minority willing to endorse political violence, and a majority expecting further rises in the incidence of such violence. The risk of self-fulfilling prophecies, whereby anticipation tips over into an embrace of preemptive action, is clear.

A third lesson here concerns the institutional platforms for preventing incumbents from the misuse of their legal and coercive powers. As Judge Ellis’s experience shows, courts have scanty tools for dealing with recalcitrant offices. A Congress aligned by party with the White House is very obviously going to do nothing. And the ex post legal checks on state violence are, thanks to the Roberts Court, minimal or non-existent. The American system of separated power, which once was thought to play that role, has largely failed. It is a sign of this failure that the term “separation of powers” can be thrown about casually by judges aligned with the president who wish to eliminate statutory or constitutional checks on the office’s powers. A concept that is so promiscuous in its applications is hardly fit for purpose when it comes to vindicating democracies. 

It is important not to naturalize this state of affairs. In many other democracies, particularly ones organized around a parliamentary form of government, opposition parties have a set of legal entitlements that allow for supervision and even vetoes on self-entrenching actions. So the failure is emphatically a local one. The general problem of “disciplining states” towards democracy can be addressed, even if it never goes away. Just not by the tools supplied by Articles I through III of the federal Constitution.

It is, however, telling that there is an official actor trying, however vainly, to row against the tide of state violence in the Evanston incident: Specifically, the mayor of Evanston Daniel Biss has called for investigations into whether Illinois or federal laws were violated in the incident. He is not alone. In Durango, California, another local police chief is investigating whether ICE agents violated generally applicable criminal laws. Historically, the political valence of American decentralization has long been associated with the disproportionate political powers wielded by the American south in the antebellum period to protect slavery, and then after the Civil War to advance racial subjugation under another name. But times change. The valence of political concepts can be unstable. Relying on decentralization as a friction on the project of deepening democracy now may be a serious mistake.

In a forthcoming law review article, Dean Zach Clopton and I develop a more general account of the larger political dynamics that crystallized in the fists repeatedly pummeling someone’s head in the crisp midwestern sun of Evanston. Borrowing a term from the political theorist Chantel Mouffe, we call this “agonistic federalism” (although we shed Mouffe’s unhealthy orientation toward the theory of Carl Schmitt): Intergovernmental conflict that proceeds with no holds barred, and that aims not just at policy wins, but at a dominance of the political field tort court. Yet, as we show, both states and municipalities like Evanston have a surprisingly extensive range of regulatory, civil, and criminal tools with which to insist on the legality of federal action. 

Leaning into this federalism toolkit, to be sure, is a way of addressing the first problem—call it ‘the legality crisis’—while worsening the second—call it ‘the polarization crisis.’ The dilemma, however, might be inevitable: As someone who has been living with ICE helicopters buzzing his neighborhood for the past few weeks, I can attest directly to the sense that acquiescence in the bullying or outright unlawful deployment of state power to intimidate generally does not have a quelling effect. Looking around at my neighbors and community, I can attest that it incites and enrages. The term “agonistic” is useful in this present moment because it captures a psychological dynamic that is increasingly locked in, a dynamic that seems to be exacerbated by every move taken on each side. Addressing the legality crisis means exacerbating the polarization crisis, in part because of the failure of arbitrator institutions such as courts and network anchors. This is a dynamic from which there is no easy exit.  

How would we know that a boundary line has been crossed, and the psychological conditions for a pooled form of political rule, i.e., a species of democracy, remains on the table? I am not sure the metaphor of a single line is a useful one. But if you must use the metaphor, then the point is surely that we cannot know when the fateful step across the line is taken. If we dissolve those jointly held dispositional predicates of democracy by trying to vindicate its legal infrastructure, it will be by inches. And we will not know we have gone too far until it is far too late.