Jim Young / Reuters

The usual form for a justice who disagrees, no matter how fundamentally, with a decision of the Supreme Court is to end the opinion with the formula “I respectfully dissent.” Justice Antonin Scalia, in particularly high dudgeon, would sometimes drop the adverb. Last week, though, Justice Elena Kagan, writing for the four justices who disagreed with Chief Justice John Roberts’s opinion in Rucho v. Common Cause purporting to withdraw the Court once and for all from passing judgment on the constitutionality of partisan gerrymanders, ended thus:

Of all times to abandon the Court’s duty to declare the law, this was not the one. The practices challenged in these cases imperil our system of government. Part of the Court’s role in that system is to defend its foundations. None is more important than free and fair elections. With respect but deep sadness, I dissent.

Kagan’s occasion for sorrow is deep not only because the chief justice left a fundamental flaw in our constitutional democracy without hope of a judicial remedy, but because of the defective reasoning by which he came to that conclusion.

Roberts is clear and emphatic that extreme partisan gerrymanders are wrongs—wrongs of constitutional dimensions. In North Carolina, for example, voters split more or less down the middle, yet the congressional delegation consists of 10 Republicans and just three Democrats. The disfavored voters were treated unequally, their unequal treatment was due to their political commitments, and so they were given an unequal voice in the legislature. This violated not only the Fourteenth Amendment’s norms of equality, but the norms of the First Amendment, as well. Indeed, it was contrary to the most basic political premise of our government: that the people choose their representatives, not the other way around.

Not only the chief but other justices as well have said as much in past decisions, but hesitated to offer a remedy. All this the chief recounts in his opinion. Yet he concludes—and now, after 40 years of temporizing, hopes to conclude definitively—that these are wrongs that are beyond judicial reparation. If there are to be remedies, they must lie elsewhere: in the states or in Congress.

Why? Roberts offers an argument with three simple steps. First, he claims that some departure from equality of voice—the notion that every voter should have an equal say—is not only inevitable, but constitutionally, and as a matter of political morality, quite justifiable. Next, he admits that the case may be an extreme deprivation of a constitutional right, an extreme gerrymander—acknowledged, deliberately engineered and frankly avowed, and arranged so as to persist though many election cycles. But, he concludes, it is not possible to draw a precise line that separates an extreme gerrymander such as this one from a less extreme one, or between a less extreme gerrymander and one that is perfectly proper. If splitting North Carolina’s seats 10 to three is extreme, perhaps a nine-to-four split is, as well. But what of eight to five, or seven to six?

The courts can act, the remedy is a judicial remedy, Roberts concludes, only if the line of demarcation is “precise,” to use the word he invokes six times. If one cannot say where the precise borderline of the extreme lies, everything must be allowed. But applying that argument to other questions illustrates its absurdity. If one cannot say how many hairs a man may have to still count as bald, there are no bald men. If one cannot draw the exact line between venal and mortal sin, then there is no mortal sin.

This is not a logical principle, and it is certainly not a juridical one. The law is replete with conclusions that depend on judgments that something is extreme, unusual, or outrageous, without offering the precise boundaries of that judgment. Take two examples from the chief justice’s own corpus. In his dissent in the 2012 case of Armour v. Indianapolis, he wrote:

Our precedents do not ask for much from government in this area [of property taxation]—only “rough equality in tax treatment” … We give great leeway to taxing authorities in this area, for good and sufficient reasons. But every generation or so a case comes along when this Court needs to say enough is enough, if the Equal Protection Clause is to retain any force in this context.”

And, similarly, in the 2007 case of Federal Election Commission v. Wisconsin Right To Life, he argued:

[Our prior precedents] … applied [the] … interest in combatting corruption or the appearance of corruption … to ads that were the “functional equivalent” of express advocacy. But to justify regulation of WRTL’s ads, this interest must be stretched yet another step to ads that are not the functional equivalent of express advocacy. Enough is enough. Issue ads like WRTL’s are by no means equivalent to contributions, and the quid-pro-quo corruption interest cannot justify regulating them.

Enough is enough, indeed.

But, as Kagan argues in her dissent, this case should be easier than many in which gathering outrage tips over into judicial condemnation. This issue has firmer bones than that.

Roberts’s faulty reasoning is built on the premise that pure equality of voice cannot and need not be demanded. And in some ways, he is correct. Proportional representation may be a possible end of apportioning representatives but it is not a necessary one, and the Court has no business making it one in the name of the Constitution. An apportioning legislature may take into account many legitimate ideals in drawing district lines: political or other historic boundaries, economic commonalities, even the desire not to pit incumbents against one another.

Drawing new lines, in other words, allows many legitimate political judgments—but disadvantaging racial minorities is not one of them, and neither should be freezing one party in and the other party out, no matter what their voting strength. Sometimes politicians are brazen enough to declare that one or the other of these forbidden ends is what they are striving to pursue, but usually they are not. They just do it.

In such a case, you can still smoke out the legitimate from the illegitimate. As the mathematician and genome scientist Eric Lander argued in an amicus brief on this issue, there are literally thousands of ways of drawing equipopulous legislative districts for any given state, arrived at either randomly or pursuant to some congeries of legitimate criteria. If only legitimate criteria are used, then political preferences are likely to be more or less evenly distributed—a normal distribution or bell curve. The farther out on the curve you get, the more likely something fishy is going on. It could be chance, or it could be the influence of some unsuspected factor, but if you are way out on the tail—as in North Carolina—you are staring at extreme partisanship, and you should be able to call it like it is and condemn it. Enough is enough.

Ah, but how small, how far out on the tail? This is not a question unique to gerrymandering; in other fields, we have settled upon practical answers and used them in practical ways: In medicine, in approving new drugs, in statistics generally, a 95 percent confidence level is usually thought to be quite enough to show causation. One hundred percent is absolute certainty and 50 is mere random chance, no causation at all. But why not set the standard at something closer to 100 percent? Aristotle explained in the Nicomachean Ethics:

It is the mark of an educated mind to expect only that amount of exactness which the nature of the particular subject admits. It is equally unreasonable to accept merely probable conclusions from a mathematician and to demand strict demonstration from an orator.

In a myriad of less consequential decisions, Roberts has shown himself to be an eminently reasonable person, an “educated mind.” But in Rucho he leads the sticklers to produce a hyper-partisan political result.

Nor should anyone find consolation in the chief’s profession of judicial modesty and abstinence, his insistence that this is just a job for other organs of government—legislatures, the states. How can you rely on gerrymandered legislatures to remedy gerrymandering? Remember, it is Roberts who wrote the devastating decision in Shelby County v. Holder, declaring unconstitutional Section 4 of the recently reenacted Voting Rights Act, thereby unleashing an immediate, utterly predictable, and devastating wave of partisan voting-rights suppression all over the South. And it was Roberts who wrote for four dissenters who found that Arizona’s independent redistricting commission was unconstitutional based on a hyper-literal reading of the Constitution’s elections clause. But for the lack of a single vote (which today he probably has), he would have closed off one of the few ways to break the vicious circle of gerrymandering.

Rucho was indeed an occasion of deep sorrow—for the Court and its chief, for the rational development of doctrine, but most of all for American democracy. Read it and weep.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.