Do you want to stop Donald Trump from rounding up Mexicans into camps? Try this: Encourage your idealistic, third-party-voting progressive and libertarian friends to drop their fantasies of an ideal, radically revised political and economic order and fight instead to protect what we’ve got. It’s the prudent thing to do, and it’s the principled move.

In a profound and persuasive new book, The Tyranny of the Ideal: Justice in a Diverse Society, the political philosopher Gerald Gaus shows that visions of political perfection are bound to lead us astray. Gaus’s argument is forbiddingly technical, but it’s not merely academic. It matters a great deal to the way we think about practical policy advocacy and presidential elections. And if your political identity is built around a dream of an ideally just society, Gaus’s argument is shattering.

Libertarian and progressive purists planning to vote for Gary Johnson, the Libertarian Party candidate, or Jill Stein, the socialist Green Party candidate, don’t think they’re pitting the perfect against good, because they don’t see much good in the status quo American political and economic order or in the morally compromised establishment’s perfect embodiment, Hillary Clinton.

So does it really matter if they draw more votes away from Hillary Clinton than from Donald Trump? What’s the big moral difference, they may ask, between a transparent thug and the bland administrator of a thuggish imperial state?

In The Tyranny of the Ideal, Gaus helps us understand why captivating theories of political perfection can make us miss the value of what we’ve got, why they tempt us to make things worse in order to make them better, and why this is so dangerous. Gaus shows that your theory of the ideal social system, whatever it may be, is almost certainly wrong. And even if it isn’t, it probably can’t work as a useful guide to political decision-making — especially not in a diverse liberal society, like ours, rife with disagreement about ideals.

Ditching our utopias for an appreciation of what Gaus, following Karl Popper, calls an “Open Society” of liberal pluralism, mutual accommodation, and incremental democratic reform brings clarity and gravity to this election season’s big choice.

Maybe we don’t need political ideals at all

In his 2009 book The Idea of Justice, Amartya Sen, a political theorist and Nobel Prize–winning economist, argued that there’s no satisfactory way to decide among several equally compelling but incompatible ideals of justice. This sort of obdurate pluralism is huge problem if you want to use an ideal to generate a picture of the perfectly just society by which to orient and guide political reform and public policy. If, say, three equally good theories deliver conflicting advice, what do you do?

Sen’s insight was that a plurality of ideals isn’t really a problem because we don’t need ideals at all. Do you really need a theory of justice to know that the boot on your neck is unjust? You don’t. As long as we can reliably identify injustice, we can focus on fixing injustices that enough of us agree to be urgent and obvious, and simply make things progressively better. That’s all we need.

If you’re climbing a mountain, you don’t need to know what the peak looks like to know what to do next. All you have to do is make sure your next step is up. Similarly, by rectifying injustices, one after another, we can just keep gaining in moral altitude. When there’s nothing left to be done, we’ve made it. We don’t need a picture of the top to get there.

Philosophers who specialize in theorizing about perfect justice weren’t impressed with Sen’s argument. The problem, they noted, is that Sen’s “just go up” model will only get you to the peak of the mountain you’re already standing on.

It may be true that we don’t need to know where Everest is to know which way is up — that we don’t need to know what perfect justice looks like to recognize and address injustice. But why think we’re standing on a tall mountain, much less the tallest one?

That is why you need a theory of the ideally just society. You need it to tell you the relative height of different mountains, of different political arrangements — and to identify which political order is the best of all. The “just go up” rule risks stranding us at a “local optimum” of justice, lacking in moral ambition. But morality demands that we don’t stop short. We’re duty-bound to shoot for the best possible social order — the “global optimum.”

Gaus argues that Sen is more right than wrong. Once we understand what it really means to commit to the optimizing perspective of an ideal, we’ll see that “just go up” is better

Ideals commit us to making things worse to make them better

Gaus argues that if we reject Sen’s climbing model out of fear of getting stranded on a morally middling peak, then our theory of the ideal will sometimes demand that we forget about correcting the injustices that are right in front of us.

Indeed, a theory-driven ideal of perfect justice is likely to demand that we go down the mountain — making our society less just — in order to set out for a higher peak. It’s inherently risky. It’s easy to go downhill, making life more harsh, oppressive, and unfair for some of us, in the hopes of eventual significant improvement for all (or almost all). But there’s no guarantee we’ll ever get to that higher mountain, to the more perfectly just society. There’s no guarantee it’s even there.

This is by no means a theoretical worry. Revolutionary communists understood the implications of aspiring to a “global optimum” and didn’t flinch. They fought the incremental reforms that led to European social democracy precisely because they knew these local improvements in justice did make things better. And they worried, reasonably enough, that this would leave workers obliviously satisfied with a less bad but still profoundly wicked system of structural inequality.

Lenin was a strategic genius who understood that a commitment to an ideal of perfect justice is a commitment to making things worse to make them better or it’s no commitment at all. His style of thinking prevailed in many places, and tens of millions of people were killed, disappeared, and starved to death. The socialist paradise never materialized. It was never even possible.

Drawing on formal models inspired by Sen, Gaus demonstrates that the practical logic of transformative political ideals forces what he calls “the choice.” Either give up on seeking guidance from ideals and embrace “just go up,” or accept that the quest for perfect justice is a risky bet that moral benefits will someday compensate for clear, immediate moral costs. If you’re willing to gamble, you’d better be very certain about the correctness of your ideal.

The thing is, no matter how confident you are, you’re probably wrong.

Pure political ideals are predictions based on guesswork

I used to think that justice requires a society of maximum individual liberty and entirely non-coercive social relations, which means there can’t be a state. My vision of the maximally free, anarcho-capitalist society seemed to me an indispensable guiding light, a compass by which to steer a course through the rugged landscape of real-world politics toward the promised land — the ideally good and morally well-ordered society.

In order for my ideal to play this orienting function, I needed to think that it was both possible and awesome. So I did. And that meant pretending to have good answers to hard questions I really didn’t really have evidence to back up.

For example, did I really know that an advanced economy can function without a tax-financed public infrastructure? Did I really know that competing private defense agencies wouldn’t just collude and establish a new state more oppressive and coercive than the one we’ve got? I did not.

Indeed, the historical evidence strongly suggests that they would. But I held out for a good long while by very tenaciously inflating the credibility of every scrap of evidence for my worldview, and mercilessly discounting the credibility of contrary evidence. This is what you do when you’re in the grip of a picture of political perfection, and it doesn’t matter what the picture is.

All our evidence about the way complex societies function comes from actual societies, past or present. Modest changes to policy in familiar institutional settings often have consequences nobody predicted or wanted.

Gaus emphasizes that the more exotic the social and institutional arrangements of your ideal, the lower the odds that it will function in the real world as it does in your imagination. But if you can’t be confident about the way a social order will actually play out, you can’t be confident that you’ve ranked it correctly relative to the other possibilities — no matter what moral criterion you’re using to do the rankings.

This isn’t a conservative point about the prudence of cautious incrementalism. This is a point about when we’re entitled to believe that a possible social system is the ideal according to our own standards. Gaus wants you to see that you can’t justifiably rank a possible system of norms and institutions as the best one if you’re just guessing about the pattern of social relations the system would sustain.

After all, if you don’t really know what it would look like, how can you compare it to the alternatives along the dimensions of evaluation that determine bestness according to your own theory of justice?

The more our social world would have to change to get to a never-yet-tried social world, the more likely you are to be wrong about it being the best. This means that a speculative utopian vision is about as usefully orienting as a demagnetized compass.

The Open Society wrings justice from diversity

Of course, we do not and never will agree about what makes a society morally best. Donald Trump, Hillary Clinton, Gary Johnson, and Jill Stein can’t all be right, and none of them is about to change his or her mind.

This fragmentation of moral opinion, characteristic of free societies, makes it impossible to arrive at consensus about the ultimate ideal, because some degree of consensus is necessary in order to undertake a remaking of society. That makes it unlikely that we’re about to climb down off this mountain on a collective expedition to some faction’s harebrained idea of the “global optimum.” If Gaus is right, that’s good.

But it doesn’t mean we’re simply stuck at our current moral elevation. Gaus shows that pluralism and disagreement improve our odds of stumbling on the truth about the good society, even if the multiplication of moral perspectives makes it impractical and objectionable to restructure society on the plan of any one sect’s ideal. This enhancement of moral vision leaves the bickering denizens of the Open Society well-equipped to recognize and remedy injustice without any overarching vision in mind.

A society of unruly pluralism and grudging mutual toleration hammering out one bitterly negotiated compromise after another isn’t anybody’s idea of perfection. But Gaus proves, as nearly as it’s possible to prove anything in political theory, that the non-ideal ideal of the Open Society is our best bet for doing progressively better.

Just go up

So we’re back where we started, with Amartya Sen’s advice to address the injustices that immediately confront us and just keep going up — or to at least keep ourselves from slipping down.

But if we can’t shake free from our fanciful political ideals, it can be hard to tell up from down. Dreams of paradise leave us comparing where we are with something we imagine to be vastly better. Compared to a pristine ideal of perfect justice, the status quo is bound to look profoundly wicked, and perhaps irremediably so, which might suggest that things will have to get worse before they can get better — that down is, in some sense, up.

That brings us back to the moral choices facing us in the current presidential election. To those entranced by a vision of utopia, the options may seem insignificant. Again we are confronted with the question: What’s the big difference, really, between a racist authoritarian thug and a hawkish imperialist technocrat? What’s worth saving in a comprehensively rigged, thoroughly unjust system?

But the truth is that our system is not so thoroughly unjust. And it is the nature of that truth that accounts for the difference between the thug and the technocrat. We are, more or less, an Open Society of diversity, mutual toleration, and free inquiry.

That’s why we have managed to gain altitude in the climb toward greater justice. That’s why, if we’re to rise higher still, it’s imperative to defend the openness we’ve got. But we can’t do that if we fail to recognize that leaders who are openly hostile to diversity and liberal toleration pose a special threat to the Open Society, and demand a special response.

From the improbably lofty height of a functional liberal democracy, the path of least resistance is definitely down. On the path up our mountain we push, always, an immense boulder, and it takes a monumental collective effort simply to hold it in place. We Americans do have an exceptional track record of upward progress, of recovery from slips.

But once the boulder starts rolling backward, it’s not easy to stop.

The fact that Donald Trump became the Republican Party’s nominee for president by promising religious and racial oppression, by expressing bald authoritarian contempt for liberal tolerance and constitutional constraint, means that our traction has already slipped. We’re already moving down the mountain. From the perspective of the Open Society, something awful has already happened.

If radicals for liberty and equality can’t be bothered to stop planning their trips to paradise with Gary Johnson and Jill Stein, if they don’t see the point of lining up behind this damn boulder and pushing like hell, we’ve already lost more ground than we know.

Op-ed by Will Wilkinson; originally run in Vox