It’s not often that opinions written by Justices Neil Gorsuch, Ketanji Brown Jackson, and Samuel Alito agree on how to decide a Second Amendment case. In United States v. Hemani they did, though each took a slightly different road to get there. The majority looked to the justifications behind Founding-era historical analogues. Jackson would have used means-end scrutiny. Alito would have said as little as possible beyond the government’s failure to back up its claims on the record. But beneath the disagreement about method lay a shared instinct about substance. The government cannot disarm a person simply by naming a permissible reason. A government could do that in any case to defeat a right. What it must do is prove that the permissible reason it names is the reason it acts on, and to tie that reason to the person it seeks to disarm.
Read more A new wave of election cases
Here, the government failed to do that with respect to Ali Hemani. A future case, built on a different record that connects a defendant’s drug use to the danger the relevant history recognizes, could come out the other way. That is not a loophole. It is the court implementing constitutional rights at their best: a demand for the right reason, proven on the facts.
How the case arose
The facts are spare. Federal agents searched Hemani’s family home in 2022 on a suspicion of terrorism-related activity. They never charged terrorism. Hemani cooperated. He surrendered a gun, pointed agents to marijuana on the property, and told them he used marijuana about every other day. More than six months later, the government indicted him under 18 U.S.C. § 922(g)(3) for possessing a firearm while an “unlawful user” of a controlled substance. It rested the charge on his marijuana use alone. For that, he faced up to 15 years in prison and disarmament for life.
Under the court’s precedent in New York State Rifle & Pistol Association Inc. v. Bruen and United States v. Rahimi, the Second Amendment analysis runs in two steps. First, a court asks whether the Second Amendment’s text covers the conduct. If it does, the Constitution presumptively protects it, and the government then bears the burden of showing the regulation at issue is consistent with the nation’s historical tradition of firearm regulation. To do so, the government must look to the history of firearm regulation. The court has said two features are important in this historical analysis: both the “why” and the “how.” More about how those relate to each other below.
The government conceded that § 922(g)(3) burdens protected conduct (being the right to possess a firearm). So it set out to carry the historical burden, and the version of the statute it defended was a sweeping one. On its reading, the statute disarms a person the moment he becomes an unlawful user of any controlled substance, and it does so until he stops. The amount does not matter. The substance does not matter. Whether the government has demonstrated that the person is dangerous does not matter. To carry its burden, the government offered one tradition pointing to a historical “how”: laws that once governed the “habitual drunkard.” That type of law provided two reasons the government said justified its action. The first “why” was the need to protect the public from “unusually dangerous” individuals who commit “violent crime[s].” The second was that the government is permitted to disarm any groups that regularly use intoxicants.
The Supreme Court concluded that the government failed to prove its action was tied to the first reason. And it rejected the second reason as not a permissible one. How the court tested those reasons reveals what the Second Amendment actually demands.
A failure of reasons, and of proof
To begin with, the court rightly resolved the case only as applied to Hemani. That was the disciplined choice. In an as-applied challenge, a court can demand that the government produce evidence connecting its action to a permissible reason, and it can make sure the government isn’t actually acting on excluded reasons through that evidentiary demand. That same evidentiary scrutiny generally does not translate to facial challenges (that is, a claim that the law is invalid in all cases), where the individual litigant drops out of view. By resolving things narrowly, the court left for another day the questions it could not responsibly answer here: the disarmament of addicts, of the presently intoxicated, and of users a legislature has specifically found dangerous.
Reviewing the as-applied challenge, let us start with the government’s argument that history supports a purpose of disarming any group that regularly uses intoxicants. As the court pointed out, a habitual drunkard, at the Founding and for decades after, was not simply a frequent drinker. He was someone whose intoxication left him incapacitated. Early America drank heavily, and the laws still reached only the incapacitated. Section 922(g)(3), on the government’s reading, reaches anyone who regularly uses any scheduled substance, without regard to incapacity. That is a wholly different target.
The government also asserted that its justification for prosecuting Hemani was a concern about dangerousness, and that § 922(g)(3) disarms unlawful users to protect the public from those who are unusually dangerous. The court suggested that unusual dangerousness was more promising as one of the “animating purposes” for historical regulation of firearms. However, the decisive question in an as-applied case is whether the government’s stated reason is an actual reason, supported by the evidentiary record, or an unsupported assertion. Here the government disclaimed any need to show that Hemani’s use made him unusually dangerous. It asked the court to treat regular marijuana users as categorically dangerous, in its own words, “without any further showing.” So nothing in the record connects his disarmament to the reason the tradition licenses.
The court refused to uphold the government’s assertions about dangerousness on the government’s mere “say-so,” without “any further showing.” The court’s own reservations name the gap. It left open “whether the government could bring a prosecution” backed by “individualized proof” about the defendant’s dangerousness, or one resting on proof that “a certain drug always renders its users dangerous.” The government chose not to supply either of these factual nexuses.
The point is not academic, as the court’s hypotheticals show. On the government’s theory, the same rule reaches the husband who regularly takes his wife’s Ambien to sleep and the student who uses a friend’s Adderall to study. The drug “makes no difference,” the court observed, nor does “how much an individual uses or the effects it has on him.” As the court put it, “to state the analogy is to expose its deficiency.” Without a factual nexus to dangerousness, the government could prevail against people like these too, on nothing more than regular usage of those controlled substances. The reason would do no work. The bare fact of use would carry the case. That is the result the court wisely declined to accept.
This is also a case about pretext, and the court treated it as one. Even assuming that disarming those shown to be dangerous is a permissible reason, the court doubted that dangerousness was the reason the government was actually acting on here. Section 922(g)(3) borrows its triggering category from the Controlled Substances Act, a statute whose schedules are populated for reasons of public health and drug control that often have nothing to do with violence. And the government’s own conduct cut against its asserted reason. It curtailed marijuana enforcement, reclassified some marijuana products, and watched most states legalize this drug. Having helped build that landscape, the government is, as the court put it, “awkwardly positioned” to call the millions who use marijuana categorically dangerous.
I have written that pretext analysis of this kind is necessary to make sure the government is actually acting on permissible reasons in as-applied settings. The 2015 case of Holt v. Hobbs is a good illustration of why. There, a Muslim prisoner sought to wear a short beard for religious reasons, and Arkansas offered prison security as its reason for refusing. But the state allowed the same beard for medical reasons and already searched prisoners’ hair and clothing. It regulated the religious beard while tolerating the identical risk everywhere else. That is the tell. If pointing to a permissible reason like prison security were enough, the government would always win. But it is not enough: The government must prove the connection between its action and that reason with evidence, and Arkansas could not. Hemani runs the same test. The reason the government pressed was not the reason it acted on elsewhere.
Read more Cameras in the courtroom?
Put these threads together and a single method emerges. A limited number of permissible reasons fixed by history. A presumptively as-applied posture. A demand that the government factually connect its action to that reason. And a check against pretext.
This is not freewheeling judicial balancing of different interests, and it is not wooden analogue-hunting either. It is disciplined scrutiny, grounded in history and constrained by the evidentiary record. It is precisely the analysis I have argued for elsewhere. Without this combination, the court warned, a “broad power to designate any group as dangerous” would let the government “quickly swallow” the Second Amendment.
A right is a constraint on reasons
We sometimes picture a constitutional right as creating an impenetrable wall. On one side sits protected conduct the government may not touch. On the other sits everything it may regulate. Some categorical rights do operate this way. But most don’t. Generally, a right operates as a constraint on the reasons the government may act on. It does not forbid the regulation. It forbids the wrong justification for it. In other words, it is a wall with certain permitted gates of entry.
As I explain in greater length elsewhere, a constitutional right is a protected reason. It gives the government a first-order reason to protect the interest the Constitution names, here the interest in keeping arms for self-defense. And it gives a second-order reason that excludes certain justifications for interfering with that interest. The right is the giving of reasons by the Constitution’s makers to the government both to protect a specific interest identified in the Constitution, and to exclude certain additional reasons for interfering with that interest.
It is worth noting how the court’s Second Amendment doctrine has moved in this direction since Bruen. There, the court asked whether history furnished an analogue for a modern gun law. Rahimi asked something subtly different. It asked whether the government had a “permissible reason” for disarming or restricting certain firearm uses, rooted in history. That shift is an important and positive change in trajectory that Hemani continues, and it tracks how constitutional rights ought to work.
Read this way, the court’s historical inquiry is doing a specific kind of work. The court frames its test as the “why” and the “how.” But in Hemani the “how” really serves the “why.” The historical analogues, whom the drunkard laws reached and how they operated, are not the object of the inquiry. They are the instrument. The court studies how the old laws worked in order to identify the reason they served. That reason was incapacitation. The drunkard laws, the court explained, reached those whose drinking “rendered them practically incapacitated and incapable of managing their affairs.” They did not reach the regular drinker. So the history matters because, as in Rahimi, it fixes which type of justification is permissible, not because the modern law must mirror an old one for its own sake. The government wanted to rely on the justification of its need to disarm any regular user of controlled substances. The history pointed to disarming a regular user only if such use incapacitated that individual. Those are different reasons, and that difference decided the case.
The concurrences
The separate writings point in the same direction from different angles.
Alito, joined by Justice Elena Kagan, would decide the case on the single ground that the government’s analogues reach only the incapacitated, and say no more. He catalogues what the record does not show: “how much he used, the strength of the marijuana he used, how many times he used it … or the degree to which this use affected his ability to exercise judgment.”
Jackson, joined by Sotomayor, would replace Bruen with means-end scrutiny, faulting its history-and-tradition test as unworkable. She reads the majority as implying that § 922(g)(3)’s “operation is not sufficiently tailored to the government’s stated purpose” (to disarm dangerous persons), which she calls “precisely the issue to which means-end scrutiny would direct our focus.”
She is right that the court is doing a kind of means-end scrutiny. She is wrong that this makes the historical inquiry dispensable, or that it requires judicial balancing. It is a type of means-end scrutiny that is historically grounded and evidentiarily disciplined. The history helps illuminate the permissible ends that could limit defeasible constitutional rights, and the court must then demand proof that the government’s means actually serve them.
Each opinion is circling the same question the majority answers: whether the government’s reason fits what it has done, and whether its actions are actually in service of a permissible reason.
The essence of a right
Hemani shows what can unite an otherwise fractured bench. The lesson is not that the government may never disarm dangerous or incapacitated individuals under this statute. It is that a reason does no work until the government proves it and ties it to the person before the court. A government that needs only to name a danger can manufacture one in any case, against any right. So the court demanded more. It required a permissible reason drawn from history. It required proof on the record. It required a check against pretext. That discipline is not an obstacle to legitimate regulation. It is constitutive of the right itself.
Read more Divided court bars federal district court review of non-final state-court judgments