The Authoritarianism of the Open Mind

A confession to begin with, which I shall try to keep in mind throughout. Some months ago I came across a paper in Psychonomic Bulletin & Review whose finding pleased me considerably, and the pleasure ought to have made me suspicious. The paper reported that subjects high in right-wing authoritarianism were worse than other subjects at correcting their factual errors — even errors of the most apolitical kind, like the persistent superstition that adding salt to a pot of water makes it boil faster — and that the deficit was worse a week later, when whatever pressure the laboratory had exerted in the moment had safely receded. I read this and I was pleased, and I filed it away to be deployed against future irritants. This is, I take it, more or less what the paper itself describes its high-RWA subjects doing — receiving a finding that fits, holding onto it past the point of verification, letting consolidation do the work that scrutiny ought to have done. I noted the irony, but not as much as I should have, because the satisfaction was already at work.

I report this small episode because it is the experience out of which the present essay was, regrettably, born; and because I want, by acknowledging it now, to keep some hold on the difficulty that an honest treatment of the subject must face. What follows is an argument I half believe and half mistrust. I shall try to mark, where I can, the places where the mistrust is loudest. The reader is entitled to mistrust me twice over, and I encourage him.

I. The brain was never designed for truth

Begin, as one must, with the species. The human animal’s belief-forming apparatus was not engineered, by natural selection or any other agency, to track reality. It was engineered to keep its possessor alive in a small group of close kin and to retain that group’s good opinion. The two purposes are not the same, and where they conflict — as they regularly and tragically do — accuracy is the first thing thrown overboard.

The evidence is by now overwhelming and reasonably depressing. Confronted with a contradiction between what they believe and what they observe, human beings do not perform a tidy revision of the offending proposition. They go to almost any length to preserve the surrounding architecture. In one recent study, more than four-fifths of subjects, when forced to revise an inconsistent belief, preferred to rewrite their general rules or invent new explanatory narratives rather than discard the small specific belief the evidence had actually contradicted. They did not update; they renovated, and the renovation conserved the original error untouched within an enlarged structure.

The picture worsens at every level of the apparatus. Good news is processed by one neural pathway, bad news by another, and the two pathways do not share equally in the work of revising one’s worldview: we are walking optimists at the level of frontal–subcortical circuitry. Add to this Dan Kahan’s findings on motivated reasoning — that the most numerate citizens, those best able to interrogate a graph, show the largest partisan gaps on politically charged scientific questions — and one is forced to a conclusion no flattering evening class on critical thinking can dispel. Greater intelligence does not produce greater accuracy on any topic that touches one’s identity. It produces only more elegant excuses.

I include myself in this. I have to. The whole intellectual interest of the finding consists in its universality, and the moment I exempt the present essayist from it I have committed exactly the error the literature describes. Whatever follows, the reader should assume the writer is also somewhere in the data.

II. Some are more stubborn than others, and other things one would like to be sure of

Within the universal predicament there is variation, and the variation is not random. Some people update their beliefs more readily than others. The psychologists who study this disposition have given it the unlovely name of actively open-minded thinking, by which they mean the willingness to consider evidence against one’s existing views, the capacity to postpone closure, and the disposition to weigh objections fairly. Its opposite is what Arie Kruglanski christened the need for cognitive closure — the craving for a settled answer, for the comfortable warmth of a finished thought. People high in this need, Kruglanski showed, “seize” upon the first plausible interpretation that comes to hand and “freeze” upon it. They are not stupid and not unobservant. They merely find suspended judgement intolerable, in much the way a smoker finds an empty cigarette case intolerable, and they will scramble to fill the void with whatever ideological tobacco is locally available.

So far, so reasonable. But here I must mark the first place where my mistrust is loudest. The move I am about to make is the move from a general account (“everyone has biases of this kind”) to a specific allegation (“and they are distributed unevenly along the political spectrum, in a direction that happens to favour my side”). This is exactly the move a right-wing authoritarian makes when he announces that, on careful inspection, the universities have turned out to be staffed by ideologues. It is not invalid for being convenient; one’s convenient conclusions sometimes happen to be true. But the convenience is data. I shall try to keep it in view.

III. The easy target

It will not have escaped notice that the names of certain political tendencies recur in the literature on belief-updating with a regularity that is either revealing or embarrassing, depending on one’s prior commitments. The 2020 paper with which I began did not stumble on right-wing authoritarianism by accident. Across two preregistered samples, with stubbornness sufficient to commend itself to its subjects, the study found that the higher one scored on the standard authoritarianism battery, the worse one was at correcting one’s own factual errors. RWA also correlates, in the broader literature, with climate scepticism, with lower scores on the Cognitive Reflection Test, with higher need for closure, with confirmatory thinking, with diminished open-mindedness. The pattern is robust enough that one would have to work hard to dismiss it.

I have, like every adult, encountered the type and could spare you the description. From perfectly reasonable starting points — that experts have sometimes been wrong, that journalists have sometimes lied, that the universities have given him real provocation — he reaches the conclusion that the climate is not warming despite half a century of converging measurements, that the 2020 American election was reversed by a conspiracy involving voting machines, the long-deceased Hugo Chávez, and at one point an Italian satellite, and that the first black president of the United States was secretly born in Kenya, a proposition surviving the production of a Hawaiian birth certificate and two contemporaneous newspaper announcements. He is impervious. He has been the recognisable villain of liberal civilisation since the Dreyfus affair. The literature is unkind to him, and he has earned its unkindness.

I have just enjoyed writing that paragraph more than I should have. I want the reader to notice, because I am noticing. The pleasure of itemising the absurdities of the other side is one of the primary aesthetic temptations of polemic, and the temptation does not vanish on being identified — it merely becomes harder to act on without feeling foolish, which is some progress, but not much. Let it be conceded, then, that the high-RWA voter is a real psychological type and that his sins are real and tabulable. Let it also be conceded that the same paragraph could be written about my side, with different specifics but similar smugness, by an able writer of opposite sympathies, and that I would find his version offensive while my version delights me. This is not nothing. It may, in fact, be the whole problem.

IV. The authoritarian who reads

It is one of the small ironies of contemporary social science that the existence of a left-wing authoritarian — a person of progressive ideological commitments who exhibits the same desire to coerce, censor, and punish that has always been catalogued on the right — was, until quite recently, treated by much of the relevant literature as a kind of bigfoot. Sightings were reported; sober researchers expressed scepticism; the consensus held that no such creature existed and that those who claimed to have spotted one were probably reactionaries themselves, motivated by a wounded vanity at the long association of authoritarianism with their own side.

This consensus has now collapsed, largely through the work of Thomas Costello and his colleagues, whose 2021 paper based on more than seven thousand subjects across six samples established with admirable thoroughness that left-wing authoritarianism is real, measurable, and behaviourally consequential. The left-wing authoritarian, like his right-wing counterpart, wishes to silence those who disagree with him. He approves of political violence under the appropriate flag. He desires the punishment of dissenters and the centralisation of state power. He shares with his rightward cousin what Costello’s group described as a “shared constellation of personality traits, cognitive features, beliefs, and motivational values” — a shared heart of authoritarianism. They are, to a first approximation, the same animal in different camouflage.

But they are not, in one specific respect, alike. The left-wing authoritarian is, on average, less cognitively rigid. He scores lower on dogmatism, lower on need for cognitive closure, lower on confirmatory thinking. He is, in the technical sense, more flexible. He can update his beliefs. He can, when the situation requires, change his mind.

Now — and this is the place where I must slow down, because the temptation is precisely to speed up — there are two readings of this finding, and the essay I started out wanting to write privileges one of them in a way I am not entirely sure it has earned.

The first reading, which is the one I have been heading toward, is that flexibility-plus-authoritarianism is more dangerous than rigidity-plus-authoritarianism, because the flexibility allows the authoritarian to update his prohibitions in real time, to keep pace with shifting orthodoxies, to remain perpetually one step ahead of his targets. The cognitive virtue serves the political vice and amplifies its reach. This is a real argument. I shall make it shortly.

But the second reading — which I owe the reader and myself — is the opposite one. It runs as follows. If cognitive rigidity is what makes the right-wing authoritarian impervious to correction, then his left-wing counterpart’s flexibility is good news: it means he can in principle be argued out of his positions. He may be momentarily certain that this week’s heretic deserves the pillory, but he is, by construction, capable of revising the judgement. The cognitively rigid authoritarian, by contrast, is the one who cannot be reached. He is Stalin in his last paranoid years, killing the doctors because he has frozen onto the idea that they are killing him. The flexible authoritarian is, on this account, a problem one can in principle talk down. The rigid one is the one who shoots the talker.

I find this second reading less satisfying than the first, but I am obliged to notice that I find it less satisfying for reasons that are partly aesthetic. The first reading delivers a sharper essay. It allows me to administer a more bracing rebuke to a wider range of contemporary annoyances. The second reading, were I to follow it consistently, would oblige me to write a more boring piece in which both kinds of authoritarian are bad in roughly proportionate ways, with the rigid kind probably worse because his rigidity is harder to penetrate.

I do not, in the end, fully accept the second reading — I shall say why in a moment — but I want it on the page. It is at least as defensible as what follows, and a writer who does not say so is not entitled to his closing remarks about willingness to be wrong.

V. What flexibility actually buys

Here, then, is the case for the first reading, given as honestly as I can make it.

The right-wing authoritarian is in many ways constrained by his own rigidity. He believes what he believes; he believed it twenty years ago; he will believe it in twenty years’ time. He cannot easily redefine the boundaries of the impermissible because he cannot easily redefine anything. His tyranny, when he gets the opportunity to practise it, has the dreary predictability of a Sunday-school sermon. One knows, more or less, what he will forbid: blasphemy, sodomy, sedition, the wrong kind of dancing. The list is unattractive but it is fixed. One can plan around it. One can flee across a border and find that the border has not moved.

The cognitively flexible authoritarian suffers from no such handicap. He can update his prohibitions in real time. The book required on Monday is removed on Friday and is, by the following Monday, evidence that one is contaminated for having read it. The word that was the polite term last year is the slur this year, and the person still using last year’s polite term is not merely behind the times but actively complicit in some ongoing structural violence. The scientific finding that was settled is now contested; the finding that was contested is now settled; and the only thing one knows for certain is that the authority deciding which is which is on one’s own side, and that the side is always — by definition, by the deepest tautology of the authoritarian temperament — correct.

Robespierre understood this. Trotsky understood it; he merely had the misfortune to be on the wrong end of his own logic when Stalin understood it slightly better. The Cultural Revolution was a six-year graduate seminar in the principle that today’s commissar is tomorrow’s counter-revolutionary, and the cognitive flexibility required to keep up with the shifting line was a survival trait for which insufficient demand existed in the more lethargic dictatorships of the right.

But here I must enter the second qualification. I have just listed regimes that killed people in the millions. I am writing in a country where they do not. The argument that flexibility makes left-wing authoritarianism more dangerous, in the present, in functioning Western democracies, has to do work the historical examples cannot do for it. Stalin’s commissars were not a problem because they were cognitively flexible; they were a problem because they were Stalin’s commissars. To use them as evidence of what cognitive flexibility “permits” is to smuggle the body count of a totalitarian state into an argument about the temperament of one’s contemporary critics. I notice myself doing this. The smuggling is rhetorically effective, and I should perhaps not undo it. But I should at least confess to it.

VI. Three cases, and what they do and do not show

Let me try, then, to make the contemporary version of the argument without leaning on the morgues of the twentieth century. Consider three cases drawn from the recent governance of advanced democracies.

A handsome young prime minister, son of a more famous prime minister, who having spent rather a large portion of his early adulthood in dark makeup at fancy-dress parties, recovered sufficiently from this minor embarrassment to invoke emergency powers against a noisy but largely peaceful protest, freezing the personal bank accounts of citizens charged with no crime. A federal court, deliberating later in the absence of the original political pressure, declined to endorse his judgement.

The compassionate, telegenic young leader of a small island nation, much beloved of the international press, who during a public health emergency instructed her citizens, on camera, to regard the government as their “single source of truth” and to dismiss any contrary information from any other source. The phrase has the merit, at least, of clarity. Had it issued from a leader at the opposite end of the political spectrum, it would have generated, at a conservative estimate, a small library of academic horror.

A barrister and former Director of Public Prosecutions, now prime minister of a once-free country, whose government has lately discovered that among the urgent threats to public order is the citizen who posts the wrong opinion on social media, and which has accordingly handed down custodial sentences for tweets that would, a generation ago, have struck any English magistrate as a constitutional joke. In several well-publicised instances, the sentences for words have exceeded those imposed for the underlying physical violence.

These are not candidates for Stalin’s chair. They are something more revealing: the everyday operation of cognitive flexibility in service of an authoritarian temperament, in countries where nobody is shot. Each instance reflects a line that was inviolable in 2019 — bank accounts of citizens not charged with any crime are not to be frozen, the government is not the citizen’s epistemic guardian, prison is not for tweets — and that has been redrawn by the gentle work of a few bad years.

I owe the reader two admissions about this paragraph. The first: I selected these three. I could without difficulty have selected three from the other side. The American republican Party has, in the same period, produced sustained efforts to overturn an election by means that involved the suborning of public officials, the harassment of poll workers, and the eventual storming of a legislature; a governor of a populous state has used the apparatus of government to punish a private corporation for a speech act he disliked; a former president has called for the suspension of parts of the constitution. To list my three cases without listing these three is to compose a story by selection, and the story is doing work the data are not strictly compelled to do. Anyone sufficiently annoyed by my trio is invited to assemble his own; the materials are abundant.

The second admission: even granting that the trio is real and that the line has moved, the question of whether the movement is more sinister coming from this direction than from the other is not settled by listing the cases. It is presupposed by the listing. If the right-wing prime minister of a country I do not happen to live in announced tomorrow that the government was the citizen’s single source of truth, I would be horrified, and I would not pause to admire the consistency of his ideology with his methods. The horror would be the same. The asymmetry I am defending — that flexibility makes the progressive version worse — is not visible in any one case. It has to be argued at the level of the type, and at that level, as I have already conceded, the argument is contestable.

VII. What I do, in the end, think — and why I might be wrong

Here, then, is what I am willing to say, with the appropriate caveats around each clause.

I think Costello and his colleagues have correctly identified something real: the existence of a left-wing authoritarian temperament that the discipline previously wished away. I think this finding is important and that the wishing-away was a scandal. I think the same temperament, expressed in progressive idiom, is responsible for a meaningful fraction of the more illiberal currents in contemporary Western politics, and that pretending otherwise — out of a sentimental attachment to the idea that authoritarianism is a property of the right — has cost us something.

I think — but here with much more hesitation — that the cognitive flexibility of the left-wing authoritarian is at least a complicating factor, and possibly an aggravating one. It allows the line to move in ways the rigid authoritarianism of the right does not. It allows yesterday’s allies to be today’s enemies on a schedule the participants themselves could not have predicted six months earlier. This makes the whole apparatus harder to plan against, harder to satirise, and harder to leave by emigration, since the exit is not from a fixed regime but from a moving one.

But I am not certain about this. I am not even confident. It is possible, on a fair reading of the same evidence, that flexibility is a feature rather than a bug — that the cognitively flexible authoritarian is the one who can in principle be brought back, and that his rigid counterpart, however more decorous in a parliamentary setting, is the one who, given full power, becomes Stalin. I do not know how to settle this between us at the level of argument. I suspect it can only be settled empirically, and I suspect the experiment has not been run. The Sinclair paper measured RWA but not LWA. The Costello papers measured both but did not run a belief-updating task. The critical study — the one that takes a politically balanced set of false claims and watches both kinds of authoritarian fail to update on the ones that flatter them — has, as far as I can find, not been done. Until it is done, the question is open, and any author who tells you otherwise is selling you a conclusion before the evidence is in.

I will go further. The voice in which this essay has been written — the voice of confident judgement, of moral clarity, of a man who has worked the question out and is now telling you the answer — is itself an instance of the disease. It is the voice of someone who has identified an enemy and arranged the world to make the identification come out right. I have tried, here and there, to interrupt the voice. I have not entirely succeeded. The cadences of certainty are easier to write than the cadences of doubt, and they sound, to most readers, more authoritative; this is one of the small tragedies of prose. But I would rather end the essay by admitting the trouble than by a stirring peroration which would be, in this context, almost a parody of the failure mode under discussion.

The willingness to be wrong is not a closing flourish. It is a discipline, and it is hard, and I am not very good at it. I am, on the present evidence, slightly better at noticing when I am failing at it, which is at least a start. My quarrel with the authoritarian temperament — in its progressive as in its reactionary form — is not that it gives wrong answers. Everyone gives wrong answers. It is that it cannot bear the suspended state in which the answer is not yet known. It hates the perhaps. It hates the I-am-not-sure. It hates the slow, embarrassed, repeatedly corrected progress of a person trying to think honestly in public.

That is the temperament against which I take myself to be writing. Whether I have written against it, or merely from a more flattering version of it, the reader will have to decide. I am, on this question as on most others, the wrong person to ask.