Keyword Authoritarianism

Further Notes on Political Correctness, Data-Visualisation, and the Death of Humanistic Inquiry

This essayletter is the second of a two-part series. The first is here. If you like what you are reading, please consider subscribing. See the “Housekeeping” section at the bottom of this letter for more information about it.


You might be convinced that nothing new or insightful can be said at this point about political correctness. Or perhaps you doubt that such a thing exists at all. If so, it might be useful to move out of the domain of politics altogether, and to see what the phenomenon, real or imagined, has in common with other pressures that exert themselves on our ways of speaking.

We generally presume that the urgency of replacing, say, “Laplander” with “Sámi” —to take an example far from the typical American experience— has something to do with respect for the preferences and self-conception of the group in question. But this presumption leaves us unable to explain plainly parallel cases, in which similar corrective pressure is exerted on speakers using outdated terms, yet where the people being described are long extinct and have no self-conception at all. Thus we are sometimes told that we are wrong to pronounce the th in “Neanderthal” as a voiceless dental fricative, as in, say, “thought” rather than as in “thyme”. The rationale has something to do with the phonetics and orthography of German, pre- and post-spelling reform, but implicit in the correction one discerns the sentiment that we are getting something wrong about Neanderthals themselves, that we are retroactively wronging our brothers and sisters of the broader Homo genus.

The phenomenon extends beyond the boundaries of the human altogether. Pressure is strong to represent dinosaurs as birds rather than as reptiles, for example. Of course, it is always good to get things right, and it is true that T. rex is phylogenetically closer to a chicken than it is to a crocodile — a truth that we did not know at the dawn of modern palaeontology, and that we do know now. But the superciliousness with which the Dinosaurs are birds dogma has come to be defended in recent years has led to some straightforwardly absurd exaggerations of the poor old Tyrannical Lizard King’s true chicken nature. And these exaggerations can only be explained as instances of a phenomenon we may rightly call “taxonomic correctness”.

We can be certain that no one cares about achieving justice for the dinosaurs themselves. The Cretaceous-Paleogene Extinction Event squelched any hope of that 66 million years ago. There is therefore obviously something else going on besides concern for the well-being or the feelings of the beings whose description is being contested. This “something else”, I contend, is a communication environment that places supreme value on the social display of a superficial variety of correct knowledge, that treats the competent attainment of this knowledge like a prize in a team-based multiplayer video game, and that simultaneously undervalues forms of knowledge that are far less easy to display. The role of new information technology in the rise of this model of learning and communication is central, but more on that in due course.

It should not be at all surprising that this “something else” can often be found not only in the way we talk about palaeontology, but also in the way we talk, and avoid talking, about our social reality, its inequalities and injustices.


I have begun with palaeontology precisely because it is of such marginal importance to our overall collective well-being. Few people will be deeply offended if you acknowledge a special fondness for outdated representations of prehistoric scenes, for the gaudy murals that still stand in underfunded natural history museums, of fern-covered groves in which, anachronistically, brontosauruses, triceratops, and distinctly reptilian tyrannosaurs are all doing their thing side by side, and a plesiosaur frolics like a seal in an adjacent sea (as for me, I say: May they never get the funding to replace these timeworn treasures with interactive “educational” plasma screens! May they never correct their precious errors!).

Around the time some of these murals were being painted, Great Britain was asserting its imperial power around the world, and France was still doing what it could to maintain at least a respectable global presence. And so at a Paris vide-grenier I flip through the accumulated household detritus that one old man has decided to put into a box and bring out into the street to sell. Among his treasures I find a reproduction of a painting by the legendary Czech dinosaur artist Zdeněk Burian, the well-known Trachodon and Tyrannosaurus of 1938, in a cracked glass frame. I continue flipping and soon I land upon a trove of stationery from some colonial administrator in French Indochina. There is an envelope, postmarked Saigon, 1929.

I notice that the colonial stationery and the antiquated dinosaur scene move me in much the same way. It would be far too simplistic to say that the feeling is one of nostalgia, though that is certainly part of it. I feel a longing for a lost world, a sharp sense of the passage of time and the inevitable loss of everything we try to pin down and make stable, in documentations and records, in postmarks, models, and maps. A desire to make it stable in memory and imagination, even if it cannot and must not be stable in reality.

I should not have to clarify that I do not in fact wish to restore the French empire. I am an anti-imperialist. I will not however allow this commitment to compel me to deny the complex feeling I have just attempted to describe, or to evade the special force of defunct toponyms. Peking, Rangoon, Madras, Cathay, Ceylon, Formosa, Petrograd, Stalingrad, Yugoslavia, Saigon: these words are accretions of history, and to recall them to mind, to relish the old maps with all the dead names on them, is not to wish to restore the successive orders of history, to relive all the injustices and violence. It is sooner to bear witness to all that was lived already, to keep history alive and memorialised.

But it is only an oppressive regime operating in the present that would seek to convince you that this relishing to which I have confessed, this delectation mixed with bittersweet longing, can have no place in the memorialisation. It is only an oppressive regime, such as the one that governs current cultural attitudes about history and our proper relationship to it, that refuses to grant that there is a difference between imagination and will, that the imagination cannot be subdued to politics, and that the political health and stability of a society is much better preserved when we allow the imagination to roam freely through scenarios, through lands and centuries, that the political will would never wish to bring about or restore.



There is a difference, in other words, between political conservatism and what we might call “poetic” conservatism, and any progressive demand that the latter be expunged in the struggle against the former is both inexpedient and supremely unjust.

It is also with regard to the distinction between these two conservatisms that we may draw into clear relief the stark boundary between left liberalism and left authoritarianism. Liberalism at its best is wary of the sort of easy analogy between soul and state that descends to us from Plato’s Republic. It tends to appreciate that politics is best restrained to the well-ordering of society, an activity that is in turn best realised by leaving our individual souls well alone. Authoritarianism (not least Plato’s) imagines that in order to get society in order, one must start from the most elementary units of society, shaping, pruning, and trimming the very contents of our inner lives.

I am, I confess, a poetic conservative. It has come to seem to me that this is largely a matter of temperament, which is something a person just gets stuck with and really can’t do much about. For this reason I have learned to be patient with those who seem truly to not understand what we risk losing each time we change the name of something, the spelling of something, who seem not to understand that words are alive, and that to despatch them, too, is a sort of violence. Sometimes I become exasperated and call these people “philistines”, and I’m sincerely sorry about that. The truth is they probably just have an innate relationship to language that is different from mine, just like different people have different reactions to cilantro.

Some words are indeed so tainted by their histories that it would surely be better for all of us, perhaps even for the words themselves, if they were killed off — I am thinking of one English word in particular, originally borrowed from Spanish, the only word in all the languages I have learned that I am literally unable to utter, as if there is a neurological blockage that would prevent it from coming out if I were to try. But most words are not like that. Most words, even the ugliest ones, have more of a mixed record, and even at their worst stand as testimony to something worth remembering.

You will perhaps think I am crazy, but four years ago I nearly cried when the Académie Française, which has vested itself with the authority to make proclamations about the “correct” use of the French language, announced a drastic scaling back in the duties of the circumflex accent. This might not seem like such a great loss, especially given that there is at most a minuscule difference in pronunciation between ô and o, or â and a. But if it does not seem like a loss, this may be because you are not one of those people who experiences words as accretions of history.

Even if silent, the circumflex typically marks, in words descended from Latin, the loss of an s following the vowel. Thus the Latin magister is reduced in Old French to maistre, which in turn becomes maître. And in 2016 the Academy informed us that from now on we are to write maitre. But by what right? A word should show its scars, I say, the signs of its previous amputations. To declare that the correct version of the word is maitre is to conceal the loss marked by the circumflex. It is to fraudulently seek to pass this word off as if it had no history. Without wishing to exaggerate the importance or the power of the Académie Française, we may at least say that this is a characteristically authoritarian move: to seek to shape our present world as if it were not inherited from the past.


Language is only the beginning. Once you recognise the value of preserving and attending to the historical accretions of words, you will in the same measure come to see your relationship to history differently. You will reject the infantile terms of the false dichotomy between those who think the good old days were “better” than today, and those who conceive history as an arrow of progress. When you come to see language as a sort of thread connecting you to the past, enabling you to commune with the ancestors, you also come to see those ancestors as your equals: not better, not worse, neither more fortunate nor less, but simply and purely equal.

You are then able to receive missives from the past without either holding them up as exemplary of how we ought to live today, or judging and scolding them for their supposed moral infancy. You are able to read ancient epic poems extolling martial virtues without any thought as to whether this conflicts with your own conscientious objection to war. You are able to read ancient philosophers whom you in some direct and literal sense take to be “wrong” in every way imaginable —in their defense of slavery, for example, in their misogyny, in their conviction that only male bees use their stingers or that comets are fatty exhalations of earth— and nonetheless to commune with them as equals. You are able to read the work of morally transgressive, reactionary, and criminal authors, and to marvel at the synthesis of work and life that they achieved. You are able to see them all, no matter how scientifically incorrect or morally opprobrious, as equally human, and therefore as equal to you. You may even take delight in surveying all that they said and did that was wrong, and imagine yourself saying and doing the same. There is nothing wrong with that, for again, the imagination is distinct from the will.

The great fallacy of the overheated romantic imagination is to allow delectation —in, say, some gory description of some ancient battle— to jump the fence from the realm of the imagination and to take control of the will. This is plainly what happens with evil men such as Anders Behring Breivik, who had to get himself drunk on stories of the valour of Charles Martel and Vlad Țepeș, from back in the heroic times, in order to harness his own will into the execution of an evil deed. And it is this childish confusion of imagination and will, of what our minds delight in and of what we may hope to construct as our political reality, that animates the fools in Stormfront chat forums who make up stories about the exceptional bravery of the Vikings and seek to use these as legitimation of their demands for a “white” ethnostate carved out of the present-day United States. And in a somewhat less overtly political vein this is the absurd and pitiable confusion behind contemporary neo-pagan movements, which draw on Wikipedia-level knowledge of Bronze Age cultures to build up the fantasy that one can recover pre-industrial spirituality, with no signal loss, in the era of late capitalism.

But we should not let their foolishness convince us that the study of rune-stones, say, or of the material-cultural evidence enabling the reconstruction of neolithic astronomy, holds no interest for respectable people. We should not convince ourselves, as the Italian left has done, that someone as benign as J.R.R. Tolkien is the enemy, or, as the Argentine left has done (according to the report of a friend who was a student in Buenos Aires in the 1990s), that the teaching of Latin at universities should be abolished because that language is intrinsically a tool of ecclesiastical repression (don’t tell Spinoza).

Nor should we be afraid to feel a little frisson of the romantic imagination ourselves when we think about these things. We might even find ourselves enjoying the work of the crackpots who indulge their romantic imagination rather less cautiously. I myself, anyhow, am quite grown-up enough to feel assured that I can handle pretty much anything that comes my way. Madame Blavatsky, Julius Evola, Lyndon Larouche, Paracelsus, Gabriele D’Annunzio, Jean Genet: bring on all those weirdos! Until 2015 or so I was part of a French Facebook group called “Non-Reactionary Readers of Reactionary Writers”. Of course, even now, in 2020, there remain some odd contradictions in academia, where left thinkers, according to taste, continue to adore (or to “use”) Carl Schmitt or Martin Heidegger. But the pride taken in such reading by that Facebook group, the explicit declaration that this is what we mean to be doing, seems now to belong to a very different era.

I often think in this connection of something I read from Martin Amis years ago. He had been assigned the peculiar task of reviewing one of the films in the Child’s Play horror series, featuring a doll by the name of Chucky that comes to life and murders people. Amis felt compelled to address those readers worried that such an entertainment might cause some viewers to become as it were possessed and to act out the scenes they’ve witnessed. “It’s nothing to boast about,” Amis writes, “but there is too much going on in my head for Chucky to gain sway in there.” It has long seemed to me that a healthy society is one that helps its members to grease up the insides of their heads with whatever Chucky-resistant contents are at our disposal, while still allowing him to enter and slip around and make a harmless clamor now and then, rather than to seek to banish the very thought of him from entering.

And what goes for such supremely dumb productions of the inane entertainment industry goes a million times more for our inheritance from the successions of centuries. Keep room in your mind for imaginings of war and violence and transgression and injustice. These are what history has to give us; these are what language preserves. Make sure there’s always a lot else going on in your mind besides, and you may be fairly confident that these imaginings will never gain such a foothold as to take over the controls, so to speak, and to begin to direct the will.



Alright, that was all solved easily enough, but now we come to the real problem: how to cultivate a free imagination in an era in which all of our fleeting interests are rigorously tracked, recorded, and analysed by machines that cannot tell the difference between imagination and conviction, between irony and sincerity, and that then feed these data back to other human beings in our community, who have by now been trained up to think like machines themselves, that is, to count keywords rather than to read texts, and to map adjacencies rather than to think things through on their own.

But let us, again, take a rather large step back.

One of the more familiar examples of the vagueness of biological species boundaries is offered by seagulls. Population x is interfertile with the neighbouring population y, and y is interfertile with z, the seagull population that neighbours it to the opposite side. But the ability to successfully mate proves not to be transitive, as the x seagulls are not interfertile with the z seagulls. The birds thus encircle the entire planet, neighbours mating with neighbours, but unable to mate with neighbours’ neighbours. If a species is understood as a group of potentially interfertile individuals, where then are the boundaries of the gull kind?

I am not especially interested in seagulls, at least not today, but this example (originally discussed by Ernst Mayr in the 1940s) may help us to make sense of a peculiar social phenomenon in the age of the internet. These days, what we might call “arguments from adjacency” are everywhere.

Suppose, for example, that online personality x opposes police brutality, supports strong environmental protections, and opposes mandatory “anti-racism” training for federal employees. Personality y opposes police brutality and mandatory “anti-racism” training, but thinks climate-change fears are exaggerated, plants and animals are of no moral concern, Native sovereignty claims on unexploited regions of the American West are unfounded, and the discovery and exploitation of untapped fossil-fuel reserves are an inevitable element of American economic prosperity and independence. Online personality z, as it happens, hates “anti-racism” training, loves fossil fuel, and thinks Officer Derek Chauvin did nothing wrong.

Now, personality x and personality z have at least something in common. But are their thought-worlds so to speak interfertile?

Today it is as if the classifiers simply assume that two people are of the same “species” on the basis of a quick inspection of a few traits, the shape of their beaks or their feet as it were, without sticking around long enough to see whether they actually mate. The science in question, not ornithology now but rather humanistic science, has outsourced most of its observational work to artificial intelligence, and is heavily reliant on data-visualisation. A more adequate human science, reliant on actual field work, human participation, conscious of the need for the heavy work of interpretation, would carve the world up very differently. But in the age of the internet, human science is fast slipping away, and historical memory is so short and so clouded that most are completely unaware of the loss.

In academia, the loss is marked by a sudden, radical lurch in what were once the humanities towards a new focus on projects of distant and machine-aided “reading” (i.e., keyword-counting, now including, with advances in AI, what you might call “key-phrase”-counting). This is a field about which I myself was once very excited. When I first read Franco Moretti’s Graphs, Maps, Trees of 2007, it was a revelation, I felt as if I was witnessing a new Renaissance. This feeling began to change around 2015 — not coincidentally, around the same moment I began to hear talk of things such as Cambridge Analytica. I still consider Moretti’s work inspired, but I attribute most of this quality to the fact that he has kept one foot in the ancien régime, the regime that trained him up and shaped his intellectual character.

Today, however, we see many young, precariously employed scholars turning to digital humanities simply because that is where the money is. I’ve been noticing what I would call the “ERC style” in philosophy (the acronym stands for “European Research Council”), in which junior philosophers in Europe, without permanent positions, are increasingly adopting an approach to philosophy that they know the data-driven and natural-science-focused ERC will fund. The result, to be blunt, is a bunch of thoughtless bibliometry. I am seldom one to police the boundaries of my discipline, but I am pretty confident in saying that what they are doing is not philosophy. Yet they get praised for their successful fundraising simply because there is so much money involved. This is how entire fields and traditions die: not by conceptual revolutions, but by the megrims of the financial backers. In the present case, government agencies are fast reorganising as the arrière-garde of trends emerging from Silicon Valley.

Another thing happened in 2015: I started noticing the proliferation of “political compass” memes, where extremely online young people were —sometimes sincerely, sometimes ironically— plotting their own or others’ political views on a Cartesian coordinate system, where the x axis represents the left-right spectrum, and the y axis represents the continuum from authoritarianism at the top to libertarianism at the bottom. These memes were mirrored in a regular feature at the back of New York Magazine, which similarly plotted fragments of celebrity gossip and related cultural bullshit on a Cartesian plane with axes connecting such contrasting pairs as “cool” and “hot”.

The culture was moving towards data-visualisation as the primary mode of engagement with the world, and, importantly, as the replacement of what was once called “criticism”. And of course this transformation extended to politics too, where we now take Nate Silver’s various tickers and graphics on his 538 page and related data-trackers to be the nec-plus-ultra of “analysis”. But this is not analysis, any more than a machine-generated keyword concordance to the collected works of Descartes is a reading of Descartes.

Silverian data-crunching began to pass for political analysis in 2016, notwithstanding the fact, recall, that all that data-crunching was still not enough to prevent him from being dead wrong about the outcome of the presidential election; haruspicy would have done just as well. But it was only with the pandemic of 2020 that data made the leap into everyday reality, and the visualisation of data came to fill the hole in our lost ability to speak about ourselves. It was now a moral imperative to “flatten the curve,” we were told, as if human social existence itself were a two-dimensional graph.


Adjacency can easily be determined by AI. A machine can tell you how often, and at what distance, the word “immortal” occurs near the word “soul” in Descartes’s Oeuvres complètes. And a machine can also tell you with what frequency people on social media who mention Tolkien also mention Breivik, with what frequency Americans who use the condemned phrase “not racist” vote for Trump, and with what likelihood “white men” who write about political correctness on the internet will wish to be delivered “content” concerning the trials and travails of Jordan B. Peterson. As to this latter example, I know it from personal experience: no matter how careful I am to articulate what I take to be a substantive, original, and well-argued point of view, the machine just keeps sending me down the same algorithmic path into the company of people I take to be, for the most part, unsavoury, frivolous, or both. I am a poetic conservative, not a political conservative. The machine can’t tell the difference.

Under such circumstances, it is not at all surprising that most good people will allow themselves to be nudged into avoiding all those keywords that set them down the wrong algorithmic path into the company of the wrong sort of people. We might call this “keyword authoritarianism”: our language, our thought, our innermost selves are being worked on, pruned, and sculpted by language-processing machines.

AI can detect engagement — it can, e.g., track me when I visit the Stormfront website to see what those fools are up to. It is much less adept at detecting the mode of engagement, e.g., ironic as opposed to sincere, or legitimate research as opposed to ideological like-mindedness. Nor can AI capture that “too much” of which Martin Amis speaks: all that other stuff that is going on in some of our heads when we Google Julius Evola that protects us from any risk of Nazi occultism gaining any sway in there.

I believe that the extreme normative pressures that many people claim to be experiencing on their manner of expression, on how they speak and on what may be said, are very real, and complaints about them should be taken seriously. I also believe that we should try to take a wide-focused perspective on what is happening, one that seeks at least a part of the cause in the tremendous transformations in our technological and information landscape over the past, say, two decades.

It is my view that one of the principal causes of the recent spike in what we experience as political correctness is the outsourcing of the mapping of our intellectual landscape to machines. That is, it is not just the authoritarian turn of malicious and ill-willed agents on the political scene, but, much more troublingly, that the pseudo-public sphere of the internet has authoritarianism built into it as the inevitable endpoint of interactions that happen there. History certainly shows us that human beings can turn to authoritarianism on their own. But the present moment shows us that abandonment of communicative good faith and humanistic inquiry to machines cannot but lead to authoritarianism.



I do not recall who it was, but I have a distinct memory from graduate school of a philosopher of mind making the casual observation that he shared 99% of his beliefs with Adolf Hitler. The professor relished the ripple of shock and exasperation that moved through the classroom full of graduate students for a few seconds, and then added: “For example, the belief that if I wish to open a door I must turn the knob, the belief that if I open my hand the pen I am holding will fall to the ground…”

This was pedagogically effective, funny, and true all at once. It was also an expression of the species of humanistic inquiry I am so afraid is being lost: the kind that recognises how much we share even with the most evil among us; the kind that the new generation of dogmatic scolds could not possibly appreciate, who emphatically do not wish to be reminded how much they have in common with their worst enemies; and the kind that no algorithm could ever possibly know how to process and channel in the right direction.

This professor’s statement, if expressed today in the form of a tweet, would have found him directed down the path to the company of the Pepe avatars or their 2020 descendants. The internet is directing far, far too many people out to the margins. The internet is stoking authoritarianism, and the fault lies to no small extent with those who have pretended that thinking may be safely replaced by data, and that humanistic cultivation and inquiry may be outsourced to machines.


Now, some Housekeeping. As I wrote in my introductory essayletter, I used to publish fairly frequently in newspapers and magazines. I stopped doing this, because I grew tired of being told what words to use, what order to impose on my paragraphs, and so on. It seemed to me that readers are much better off having direct, undistorted access to what a writer has to say. I hope the last month of essayletters from me has proven this conviction correct. I do, however, miss the supplementary income that writing for newspapers and magazines once provided. Therefore, the ideal arrangement for me is the one in which my readers here, at least those in a position to do so, help me to make up for that shortfall with paid subscriptions. Thank you once again. I am very grateful for your readership.