Over the past few years, the American public has been bombarded with stories of abuses of power by police officers. Eric Garner was choked to death by the police in 2014 for selling cigarette packets without tax stamps. Philando Castile was shot seven times at point-blank range by a Minnesota police officer in 2016 for merely saying that he was carrying a legal firearm. And in March of this year Boulder police brandished their weapons at Zayd Atkinson, who was picking up litter outside of his own residence at the time.

To be sure, police brutality is often rather different than reported on the evening news. As Heather Mac Donald has laid out in The War on Cops, there is a virulent anti-police sub-culture which often twists facts in order to make it seem as though police officers are predators. But for all that, it is impossible to deny that there are instances in which helpless citizens are attacked and even killed by officers of the law for no apparent reason.

This raises a very uncomfortable question: What should we do if we witness such an attack taking place? Most of us would probably stand back and defer to the police officers’ judgment, rationalizing even the beating of an unarmed suspect as somehow necessary to the larger purpose of police work. But is that the right thing to do? If we see something that is clearly an episode of police brutality and have the means at our disposal to stop it, may we? Must we?

Georgetown University political philosopher Jason Brennan attempts to navigate these hard moral dilemmas in his new book, When All Else Fails: The Ethics of Resistance to State Injustice. Provocatively, Brennan argues that citizens can and should intervene when agents of the state are harming other citizens. If a police officer (or other government authority) is attacking an innocent person, Brennan believes that ordinary citizens have the duty and the right to fight back.

Start your day with Public Discourse

Sign up and get our daily essays sent straight to your inbox.

Brennan’s Two Foundations

Brennan rests his argument on two foundations. First, Brennan avers that “the special immunity thesis is false.” The special immunity thesis holds that anyone acting on behalf, or somehow clothed with the authority, of democratic governments “enjoy[s] a special immunity against being deceived, lied to, sabotaged, attacked, or killed in self-defense or the defense of others.” The people choose their leaders in a democracy, so if someone is working for such a government in an official capacity then he or she must, by simple logic, also be exercising legitimate authority. But for Brennan, this is specious reasoning. Just because someone is democratically deputized does not mean that he or she enjoys any more latitude to commit wrongs than another person not similarly privileged. Citizens are under no obligation to defer to wrongdoers just because they might be wearing a uniform or sporting a badge.

The second of Brennan’s two foundational arguments flows from the first. Brennan maintains that “the moral parity thesis is true,” the natural corollary to the “no special immunity” stance. Brennan argues that “We [i.e., civilians] may respond to governmental injustice however we may respond to private injustice. Government agents are due no greater moral deference when they act unjustly than private agents are due.” Right is right and wrong is wrong. If an average person had put Eric Garner in a chokehold on a New York City street, anyone would have been justified in stepping in and saving Garner’s life. This is still the case when the person who has Garner in a chokehold is in blue. Whoever the agent, an evil act is always evil and no special consideration should be given to position or rank.

At first blush, Brennan’s twin theses would appear to carry great moral weight. Who has not felt indignation, even outrage, when witnessing injustices carried out by the authorities against someone whom the law and custom prevents fighting back? The group of Los Angeles police officers whaling on a defenseless Rodney King in 1991 was a famous example of how far a civilian must, in the eyes of the law, defer to state power in order to avoid being charged with “resisting arrest.” King was prostrate on the ground and trying desperately to ward off the officers’ blows, but he did not return their assaults. The person videotaping the whole thing did nothing to try to stop the beating. Would it not have been better if someone had tried to help King?

Yes, it would have, Brennan argues. Appealing to political philosopher John Hasnas’ explication of English common-law doctrines, Brennan holds that when people act in reasonable self-defense (there are many shades of nuance here that Brennan acknowledges and skillfully explores), they are not just excused from their act of having taken another life, they are positively justified in it. It’s not merely permissible to kill someone who is trying to kidnap a child or commit rape or some other horrible crime—it’s the right thing to do.

Brennan’s Blind Spots

These are compelling arguments, and there is much in Brennan’s book with which a reasonable person can agree. On the narrow terms of many of the scenarios given, it would seem justified if an average person were to shoot a group of police officers beating an already-subdued suspect, or lie to federal agents looking for someone so they could arrest him on trumped up charges. Indeed, at the micro level, when we are talking about set situations such as those that Brennan offers in abundance, his overall argument holds up nicely. But at the macro level, it does not. When All Else Fails suffers from two fatal argumentative blind spots, one epistemological and the other having to do with unintended consequences.

To take the consequences first, while Brennan is at pains to emphasize that he is not calling for revolution (peaceful or violent) or anarchism—he is “merely argu[ing] that you may defend yourself and others from particular acts of government injustice in the same way that you may defend yourself and others from particular acts of civilian injustice”—the obvious outcome of people taking the law into their own hands would be precisely the kind of chaos that Brennan rightly disavows. No matter how neatly the argument is framed, the program laid out in this book would be a nightmare if realized.

I share Brennan’s dismay at the overreach of the state and the countless harms that government inflicts on people in the United States and around the world. But imagine if people began killing government officials in broad daylight, even when those officials were caught in flagrante delicto committing heinous crimes against innocent passersby. The best-case scenario would be that police would abandon the areas where they were being attacked—a boon for gang members but a curse for average citizens. (Crooked cops are still subject to IA investigations; Crips and Bloods are not.) The worst-case scenario would be that government officials would decide that an armed insurrection was underway and respond accordingly. One-off resistance, even when done with the best of intentions, could easily metastasize into an all-out civil war. Brennan is not calling for such an outcome, but if people start shooting police officers then we are very likely to get just that.

There is also an epistemological flaw. Like many other academic ethical theorists, Brennan’s many hypothetical scenarios are basically variations on the trolley problem. The problem is that these scenarios get human beings all wrong. The trolley problem as a thought experiment is a pointless exercise in bad anthropology, and When All Else Fails provides countless examples of why this is the case.

As Brennan points out, the trolley problem was first proposed by philosopher Philippa Foot in 1978. Everyone knows the basic set-up: a train is barreling down a track and approaching a divide. Someone is at the switch and can choose whether to send the cars down the left side, killing x people with y characteristics, or the right side, killing n people with q characteristics. Should one run over the doctor working on a cure for cancer, or the group of ten psychopaths? Should one kill the elderly musical genius or the handicapped child? Which lives are more important, in other words, and how and why do we so decide?

The trick of the trolley problem is that it is, in its essence, a little morsel of ideology. Like other ideologies, it asks you to imagine from the outset that the world is not as it really is. Once you have agreed to the suspension of reality, then you are stuck trying to figure out answers to problems that are by nature unrealistic. It is like Marxism, which first demands that we accept dialectical materialism, an absurd precondition. Or it is like liberalism, which insists that we denature human beings from human societies and posit everyone as a Hobbesian mini-sovereign operating in an atomistic world. But the human person, and the real world, are much richer than such artificial stage-lighting would suggest. The trolley problem fails to show us the true complexity of any aspect of the real world in which we live, so the conclusions that follow are almost always untenable.

Thus, trolley problem-type thought experiments fail to show that scenarios of wanton violence are often the direct result of other, buried assumptions about society. For example, Brennan takes it for granted that being against the legalization of marijuana, in favor of the criminalization of abortion, and against unrestricted immigration are irrational positions to hold. After all, the social sciences have produced reams of data showing why people who subscribe to such positions are mistaken. But this is just the trolley problem writ large. The social sciences make stark what in practice is gray. Data alone do not flesh out the whole picture of mankind, which is not a paint-by-numbers canvas. Abortion, unrestricted immigration, and drug use may look fine on paper. But when a neighborhood is hollowed out by drug abuse, broken families, unemployment, the devastating grief of having lost one’s baby, or the horrors of sex trafficking and gang activity, then the only way to maintain some semblance of law and order is to send in police officers armed to the teeth. Just before the doomsday clock strikes midnight, we are asked to decide whether we should shoot the policeman or not. This is the folly of the trolley problem. Brennan’s scenarios are snapshots of how “all else fails,” but not why.

Here’s a good example. At the end of the book, Brennan brings up Jean Valjean, whom he praises for rescuing Cosette even while “obsessive Inspector Javert is in pursuit.” True, but Valjean was not always such a good person. He was not born a hero. It took an act of senseless kindness—a good bishop covering for Valjean even though the latter had been caught in the act of stealing—to turn Valjean’s heart from selfishness to heroism. Without the bishop’s irrational gesture of complete anti-self-defense, Valjean would have continued to live in the same neo-Darwinian world of the mean streets of post-revolutionary France. Brennan spends much time discussing consent, legitimacy, and authority in the keys of Hobbes and Locke, but if it is still true that culture is upstream of politics, then, if all else has failed, it is probably at root a cultural, and not a political or procedural, solution we need.

The temptation to shoot a wayward cop for victimizing an unarmed civilian may be very real. But it is precisely by refusing this temptation and hewing to the better angels of our nature—winged not with retaliatory violence but with patience and love—that society will change. This can come only by being the good bishop, not a timely marksman. When All Else Fails is framed as a last resort. What is needed, though, is a return to first principles.