Introduction

At this year’s Republican National Convention, President Trump shocked even his own shock-proof base by demanding that members of his audience raise their right hands and take a pledge to “cheat like hell” in the upcoming midterms, stating, “I don’t care if I’m registered or not, I’m going to try and cheat like hell like they do.” While many took Trump’s remarks as a joke, it is worth remembering that this is the same President who famously called the Georgia Secretary of State in 2021 to ask him to “find” 11,780 more votes to flip the results of the 2020 presidential election in Georgia; who tried to disrupt the certification of the presidential election in Congress by egging on a crowd to “march down to the Capitol” on January 6, 2021; and who expressed regret that he did not follow through with a plan to use the U.S. Army to seize voting machines.

Trump is already claiming that Democrats cheated in the California primaries and has (unsuccessfully) tried to change U.S. Postal Service regulations regarding mail-in ballots on the grounds that the current requirements are vulnerable to cheating. His Department of Homeland Security has claimed that non-citizens were voting in Nevada—an assertion that has been thoroughly debunked. And he has accused China of meddling in the U.S. midterm elections. In this essay, we consider the worst-case scenario: an attempt by the President of the United States to use the U.S. military to influence or control elections. Possible orders in this regard range from sending troops to polls to intimidate voters, to more direct military intervention, such as declaring martial law and delaying or shutting down voting in some places altogether. 

In 2024, we surveyed the various laws that could apply to these scenarios, including 18 U.S.C. § 592, which makes it unlawful for troops to appear at polls under most circumstances, the Posse Comitatus Act (18 U.S. Code §1385), which forbids the use of federal troops for law enforcement purposes, and the Insurrection Act, a series of laws found at 10 U.S.C §§ 251-53 that give the president broad discretion over use of the military in case of a domestic insurrection. Since then, though, there have been various developments.

The Implications of Trump v. Illinois

In Trump v. Illinois, decided in December 2025, the Supreme Court put a stop to the President’s use of one particular law to federalize and deploy troops in case of domestic unrest. 10 U.S.C § 12406(3) allows the president to use federalized forces to “execute the laws,” but only when the President is “unable” to execute the laws using “regular forces.” Following a close reading of the statute, the Court determined that Trump’s use of 10 U.S.C. §12406(3) to deploy federalized troops all across the country during much of 2025 was unlawful, because the government had not shown that regular federal troops were actually unable to execute the laws without additional assistance. 

But the statute reviewed in Trump v. Illinois provides for two other circumstances that enable a president to call up a state national guard—the first in case of foreign invasion (10 U.S.C. §12406(1)), and the second in case there is a “rebellion or danger of a rebellion” (10 U.S.C. §12406(2)). In the hands of an unscrupulous, “cheating” Commander in Chief, both provisions could potentially be in play in November. Whether lawfully invoked or not, the time needed to counter such deployment orders could mean that it turns out to be too late to impede interference with the election. There are some signs that Trump is reflecting on these additional provisions in the statute. For example, he has used the term “invasion” to refer to illegal immigration, and, during the litigation in 2025, he falsely claimed that a “rebellion” was occurring in Los Angeles to justify his deployment of the California National Guard to that city.

There has been no judicial elaboration of the meaning of “invasion,” but the circuit court opinion reviewed in Trump v. Illinois does help to explain “rebellion” a bit more. The Seventh Circuit clarified that “Political opposition is not rebellion” and that “a protest [does not] become a rebellion merely because of sporadic and isolated incidents of unlawful activity or even violence committed by rogue participants in the protest.” A “rebellion” requires “deliberate, organized violence to resist governmental authority.” In view of the narrow view of the meaning of “rebellion,” any attempt to invoke the “rebellion” provision of 10 U.S.C. §12406 would likely ultimately fail, but would it fail quickly enough to avoid damage to the democratic process?

Does the Ban on Troops at Polls Limit the Insurrection Act?

The Insurrection Act has been much discussed by commentators as creating the greatest vulnerability for military involvement in U.S. elections. It provides for domestic deployment when a state requests aid to suppress an insurrection, and when the President considers it necessary to enforce laws or suppress rebellion, or to “suppress insurrection, domestic violence, unlawful combination, or conspiracy” if they hinder or obstruct execution of the laws or of civil rights. This 1807 law has only been invoked a small number of times since passage and is consequently open to significant interpretation. It has been debated, for example, whether, and if so, when a president can lawfully invoke this provision over the objection of governors. Another much debated question is whether the Act admits of interpretation by federal courts at all, in light of an 1827 decision called Martin v. Mott. The concern is that a federal court could decline to hear a challenge to a deployment of federal troops altogether, based on the argument that Martin forbids judicial review. 

Even assuming that any of the requirements for invoking the Insurrection Act were met, or that federal courts did not put a stop to an illegal invocation of the act, there is the critical question of how such an authorization of domestic military use interacts with the specific limitations created by 18 U.S.C. § 592, namely the ban on sending federal “troops or armed men [to] any place where a general or special election is held, unless such force be necessary to repel armed enemies of the United States.” Thus, even if the president has lawfully invoked the Insurrection Act, this does not mean he can then use the troops he deploys at polls. 

Traditionally, a statute passed later in time than a previous statute is interpreted to control or trump the earlier statute, and Congress last amended the statute regarding troops at polls in 1909—decades after the passage of the Insurrection Act. However, an OLC memorandum from November 4, 1968 concluded that § 592 does not limit the President’s Insurrection Act power. This memorandum—only one and a half pages long—is sparse in reasoning and relies solely on its conclusory observations regarding the legislative history of each enactment. But much needed judicial clarification may yet come prior to the election, since § 592 has been raised in a new lawsuit regarding the FBI’s plan to send agents to the polls, given that federal agents count as “armed men” under the statute. It is critically important that the District Court of the District of Columbia resolve this matter, since, as we have stressed, there may be no time for judicial intervention on November 3rd. 

Assuming the D.C. District Court is even willing to rule prior to the election on this matter, this lawsuit still may not resolve all issues relating to the potential for conflict between § 592 and the Insurrection Act. Even if the court decides that the prohibition on troops at polls takes precedence over the president’s authority under the Insurrection Act, the conflict between the two statutes is entirely avoided if the president were credibly to claim that the exception in § 592 applied—namely that deploying troops to polls was necessary to “repel armed enemies.” Might Trump take the same liberties with this language—which has received no judicial interpretation—as he did with “rebellion” in the 2025 litigation in Trump v. Illinois? It is not clear that Trump’s motivations, even if made transparent, will be relevant to the legal question. Quite clearly it would not be relevant to a determination of whether the president was violating, say, the Posse Comitatus Act, since the PCA is a criminal statute, and under the 2024 immunity decision in Trump v. United States, the president’s motive is not relevant to determining whether he is acting within the ambit of his core constitutional authority.

The Military Servicemember’s Conundrum: Obedience in the Shadow of Uncertain Legality

Let’s imagine that with all of the foregoing, the president decides to invoke the Insurrection Act and uses the pretext of “insurrection” to dispatch troops to polls, or even to shut down polling places and call a halt to voting. Or suppose he claims that elections must be halted in certain swing states because the Chinese are interfering with our elections, and then uses that claim to dispatch troops to quell resulting protests under 10 U.S.C. §12406(2). What should servicemembers who have been ordered to deploy for one of these two fatuous reasons do? Must they participate in a transparent undermining of U.S. elections, in violation of both state and federal election law? This question implicates the law of superior orders in military justice. 

Military law breaks down orders into three categories: lawful orders, unlawful orders, and patently unlawful orders. Lawful orders must be obeyed on pain of criminal sanction. On the other end of the scale, patently unlawful orders must be disobeyed. But the middle category—unlawful orders that are unlawful but not patently so—are more complicated. For this middle category, one would be excused if one obeyed and the conduct were criminal, and if one refused to obey and was prosecuted for disobedience, one could raise the illegality of the order as a defense at trial. But refusing to obey an order in this middle category runs a grave risk: the servicemember disobeys “at his own peril” since it is he who bears the cost of an erroneous assertion of illegality.

Whether or not a deployment to the polls violates § 592 will thus involve complicated legal questions that are currently unresolved and are unlikely to be resolved before the election, even if the D.C. District Court acts. These include not only questions about the interaction of the statute with the Insurrection Act, assuming troops are called forth under that authority, but also the definition of “enemies” under §592. Given this uncertainty, and given the nature of the conduct, it is almost certainly true that these orders—if illegal—will not be patently illegal. They will instead be in the middle category, where the officer will have an option to obey or disobey. Operating in this zone of legal uncertainty, while also possessing a sincere concern for the integrity of our elections, places the scrupulous servicemember in a bind: On the one hand, officers have sworn an oath to uphold and defend the Constitution; their duty to follow the law takes precedence over their duty to adhere to presidential orders. But the version of the oath of office for ordinary enlisted also contains a promise to “obey the orders of the President.” For both officers and enlisted personnel, though, their oath to uphold the Constitution means that the promise of a criminal law excuse is cold comfort. The possible distortion of concepts such as “enemies” and “rebellion” places U.S. servicemembers in a bind—it may make them pawns in Trump’s larger game and could therefore force them to be unwitting or unwilling participants in presidential “cheating” regarding the election.

Prior to 2024, a servicemember could largely rely on the fact that when a president issued an order, it would have been thoroughly vetted by the president’s lawyers as well as by members of the JAG corps before being executed. In other words, the fact that the president’s neck would also be on the line presumably led the president to rely more heavily on his legal advisors. But in view of the Court-fabricated gift of presidential immunity, no such assurance presently exists. Moreover, the JAG Corps and other lawyers in the Pentagon have been pressured, fired, and intimidated into going along with the president’s agenda and ideology. Legal advice on military action can no longer be freely given, for fear of retribution from both military and civilian leadership. Retribution of a similar nature was in evidence in the attempt to discipline Senator Mark Kelly for merely stating the law of superior orders, as the authors made clear in a brief they filed in his D.C. Circuit case. While Senator Kelly won resoundingly in district court, and is likely to prevail at the D.C. Circuit, the DoD’s actions have presumably had a chilling effect on those seeking clarification on the law of superior orders, or guidance on the legality of a specific order.

Conclusion

The question of the deployment of troops to polls, or other involvement of members of the U.S. military or federal agents in the midterm elections, is thus highly complex. If courts cannot or do not intervene, the best hope for the future of U.S. democracy may lie in the training and integrity of servicemembers, and hopefully that of FBI agents as well. As President Franklin Delano Roosevelt once declared, “Democracy cannot succeed unless those who express their choice are prepared to choose wisely.” Let’s hope they are.