I imagine many readers will find this statement of the problem with political epistemology very appealing; critiques of the pathologies of elite academic discourse play well on Substack, particularly when the basis of the critique is that academic elites are too progressive, too left wing, too woke, or whatever. So I want to annoy those readers by suggesting that we can fix these problems by adopting an approach from the sociology of scientific knowledge—the strong programme—which is unfashionable in sociology and wildly unpopular in analytic philosophy.

The Strong Programme in the Sociology of Scientific Knowledge

The strong programme is associated with the “Edinburgh school” in the sociology of scientific knowledge. Prominent members of that school include Barry Barnes, David Bloor, Donald MacKenzie, and Steven Shapin. Unless you work in science and technology studies, or some related area, you are most likely to have come across the strong programme in the context of the “science wars” (it was one of the, though not a main, targets of this book). While I’ve never tried to substantiate this, I suspect the assimilation of the strong programme to radical forms of social constructivism is a bit of a mistake. The intellectual ancestry of the strong programme lies in Hume, Wittgenstein, and a kind of suspicion of rationalistic understandings of human knowledge that has a lot of affinities with the arch anti-rationalist Michael Oakeshott, for whom knowledge was also to be understood as embedded in social practices. The links to intellectual figures with strong ties to conservative political thought (Hume and Oakeshott, but maybe also Wittgenstein) complicates the idea that the strong programme belongs with postmodern neomarxist deconstructionist critical theory, or whatever you want to call it.

The strong programme is usually defined by four methodological tenets which are set out in Bloor’s book Knowledge and Social Imagery. These tenets state that a sociology of scientific knowledge must be:

1. Causal—concerned with the conditions (psychological, social, cultural) that bring about belief.

2. Impartial—with respect to truth and falsity, rationality or irrationality, success or failure.

3. Symmetrical—the same types of cause explain true and false beliefs alike.

4. Reflexive—the programme’s own explanations must be applicable to itself

The strong programme contrasts with the “weak programme”, which sets out to provide causal explanations of false scientific belief—cases where scientists, whether individually or in collectives, accept theories that turn out to be wrong. It is entirely natural to look for causal explanations of missteps in the process of science. Perhaps a false theory gained widespread acceptance because it served certain important political purposes, or exerted some kind of cultural attraction which explains why it gained a hold over its adherents; consider, for example, Lysenkoism, phrenology, or early 20th century eugenics.

The strong programme goes beyond the weak programme because it provides symmetrical causal explanations of true and false beliefs; both must be explained via some combination of psychological, social and cultural causes. If you want to try and explain the acceptance of a false theory in terms of its cultural attraction, you should be open to explaining a true theory in the same sort of way. And it is meant to apply to itself; we can inquire into the psychological, social and cultural causes of acceptance (or rejection) of the strong programme, or of any beliefs we end up with about the causes of particular scientific beliefs that result from carrying out the programme.

These methodological tenets lead to a kind of methodological relativism. The aim is to try and understand how someone (an individual scientist) or some collective (a group of scientists) might accept a theory given some combination of psychological, cultural and social factors irrespective of whether that theory is true or not. If the relativism remains methodological, it is hard to see what could be so objectionable about it. It is common to think that false theories end up being accepted because they are in some way or other pleasing to those who accept them. But true theories can be as pleasing as false ones. Similarly, it is common to think that false theories are accepted because they fit with the cultural zeitgeist (see eugenics in the first part of the 20th century). But true theories can fit with the cultural zeitgeist too.

One reason why the strong programme was spectacularly unpopular in analytic philosophy is that it tended to blend these methodological tenets with more substantive forms of relativism and an attempt to reduce normative explanations of scientific belief to causal ones. If however we resist this temptation we are left with what seems to me to be a sensible way of going about understanding scientific knowledge as a natural phenomenon. The sociologist of scientific knowledge should look into why scientists accept the theories that they accept, focusing on the proximate causes of these acceptances. The idea is not that all scientific theories are “equally true” or “equally valid”. The idea is rather that, when we look into why a theory comes to be accepted within a community, there is typically a psychological, cultural or social explanation for its acceptance, whether that theory turns out to be true or false.

The Strong Programme in Political Epistemology

I think we need a version of the strong programme in political epistemology. Political epistemology can be defined as the study of how we come to hold political beliefs—what causes them, what justifies them—and what social conditions make certain claims count as knowledge in political contexts. A political epistemology that adopted the strong programme would follow these four tenets:

1. Causality—It would be concerned with the causal conditions (psychological, social, cultural) that bring about political belief.

2. Impartiality—It would be impartial with respect to the truth or falsity, rationality or irrationality, of political beliefs.

3. Symmetry—It would be symmetrical, that is, use the same types of cause to explain political beliefs which the theorist approves as beliefs which the theorist disapproves.

4. Reflexivity—It would be reflexive, and in particular apply itself to any political beliefs driving the theorist.

It is easiest to see what these tenets require of the political epistemologist by considering some ways in which they might be violated. You could, for example, violate the first tenet by explaining some political beliefs (most likely, the ones the theorist likes) in terms of their obviousness, by saying that they are self-evident, yet explaining other beliefs (the ones the theorist doesn’t like) in psychological terms (they satisfy some psychological need, for example) or sociological terms (perhaps they are the result of misinformation or propaganda).

Imagine a theorist who sees no need to explain why people favour redistributive policies, or are supportive of more immigration—they view these policies as self-evidently correct, and so in no need of psychological or sociological explanation—yet reaches for psychological or sociological explanations of free-market or anti-immigration beliefs. This would be a clear violation of both impartiality and symmetry: the political beliefs the theorist approves of (views as true or at least rational) are to be explained in one way (they are self-evident) whereas other political beliefs (those the theorist disapproves of) are to explained in some other way, for instance via some psychological disposition (flaw) that leads people to accept them, or via some kind of sociological mechanism.

This kind of approach to political epistemology is implicit in some very common ways of talking about political beliefs, and in some popular views in parts of the academic literature. For example:

• It is common to attribute “bad” political beliefs to psychological flaws or “epistemic vices” like closed-mindedness and arrogance. When this goes along with the idea that these flaws and vices are more common among one political tendency than others (Brexit or Trump voters, say) we have clear violations of symmetry and impartiality.

• It is common to assume in discussions of polarisation that some political groups or “tribes” (those the theorist doesn’t like) engage in identity-driven social signalling while others are simply responding to facts. The idea that the psychological mechanisms that tie belief (or the mere expression of belief) to group identity somehow only afflict the parts of the electorate that the theorist doesn’t like is also a clear violation of symmetry and impartiality.

• Many view our current political era as a “post-truth” moment, in which “misinformation” has run amok and produced all kinds of bad political beliefs. While this could be consistent with symmetry and impartiality—if, for instance, the theorist was worried that our information environment is so polluted that it renders rational engagement with evidence very difficult, if not impossible, for all of us—it is often put forward in a way that clearly violates both symmetry and impartiality. These problems are viewed as disproportionately afflicting some political tendencies, leaving others in the clear and still able to form political beliefs in the normal way, through rational engagement with evidence.

What unites all these ways of thinking about political beliefs is the idea that we need to invoke psychological, cultural or social explanations for some beliefs—the bad ones, the ones held by our political opponents—but not our own political beliefs, or the beliefs held by people and groups we approve of. This is the equivalent of the weak programme: political beliefs that are presumed to be false or irrational require psychological or sociological explanation but political beliefs that we hold, or that we deem to be true, do not.

The strong programme demands symmetry and impartiality here. If you think your political opponents only believe what they do because those beliefs play some sort of social signalling role, or that they only pretend to believe what they say they believe in order to send some kind of signal, then you should consider whether the same might be true of yourself and those with whom you align yourself politically.

While nobody—as far as I am aware—describes themselves as following the strong programme in political epistemology, some of the best work implicitly follows something like this method, albeit not always perfectly. Let me give just three examples (there are many more).

First, Neil Levy’s book Bad Beliefs argues that the beliefs of science sceptics are typically the product of rational responses to corrupted epistemic environments rather than individual irrationality. The “bad believer” forms their beliefs in the way that everyone forms their beliefs: they defer, relatively uncritically, to recognised authorities. The problem is that their environment is corrupt (those authorities have false beliefs) so the resulting beliefs are false. While Levy offers a style of explanation that is largely symmetrical, it is fair to say that his political sympathies are evident from the book—more on this below.

Second, C. Thi Nguyen's very influential work on echo chambers does a bit better than Levy, at least so far as symmetry and impartiality is concerned. He distinguishes “epistemic bubbles” (where people simply aren’t exposed to relevant information) from echo chambers (where those within the chamber actively distrust outside sources). Both bubbles and echo chambers exist across the political spectrum, and they work because the underlying mechanisms—deference to the information you are provided with and to recognised authorities within your community—are universal.

Finally, while he is a psychologist not a political epistemologist, Dan Kahan's cultural cognition programme provides a psychological framework that some of my favourite political epistemologists draw on in their work (I do too). One of his central findings is that assessments of risk track cultural and political identities across the political spectrum. Kahan is careful to point out that this includes people whose political identity is “pro-science”—this phenomenon is not confined to climate deniers or anti-vaxxers but is a general feature of how human beings process politically charged information.

Description vs. Normative Evaluation

We need a strong programme in political epistemology because, otherwise, it risks becoming an exercise in explaining why beliefs that the political epistemologist thinks are false or irrational are, in fact, false or irrational. It also risks becoming a means of pathologising political opponents and political tendencies with which the theorist disagrees. However, it may look like one of the tenets of the strong programme—impartiality—sits awkwardly with the normative ambitions of political epistemology. If we are supposed to explain true and false beliefs, rational and irrational beliefs, by the same types of cause, what room is left for saying that some political beliefs are better justified than others?

If I’m being totally honest, I’d be quite happy to do political epistemology without much in the way of normative assessment. But some people have told me that this isn’t sufficiently philosophical. So it seems I need to have a better answer than that.

One possible answer is that the strong programme provides a methodology for just one part—the descriptive part—of political epistemology. The descriptive part of political epistemology sets out to understand the formation of political belief: by what processes do people arrive at their beliefs about politics? Are those processes the same or different to the processes by which they arrive at other, non-political beliefs? What sorts of social structures play an important role in the formation of political beliefs? And so on. The strong programme provides a methodology for answering these sorts of questions in a way that doesn’t simply import assumptions about which political beliefs are true or false, which are rational or irrational.

The worry here is that, if we don’t adopt the tenets of the strong programme—if, for example, we end up with some version of the weak programme—then we will end up mixing normative evaluation with description and understanding. The theorist who begins with the assumption that climate scepticism is irrational will naturally reach for psychological explanations of it; the theorist who begins with the assumption that the climate activist is simply responding to the evidence will not feel the need to causally explain why the activist has ended up with the beliefs that support their activism. We need the strong programme, or something like it, to avoid a descriptive political epistemology that is geared towards vindicating the political assumptions and prejudices with which we began.

Once you have a genuine understanding of the causal conditions under which people form political beliefs—the psychological mechanisms, the social structures—you can then (if you want) bring normative theories of rationality, justification, and knowledge to bear on what you have found. But we don’t need a special theory of the rationality of political belief; we have theories of rationality already. What is distinctive about political epistemology is not the normative theories which we can use to evaluate political attitudes and behaviour but the attitudes and behaviour with which it is concerned.

A Varied Diet of Examples

There is a final problem, which is that even a theorist who scrupulously follows the tenets of the strong programme might still end up implicitly favouring some political tendencies, simply through the selection of examples. Levy is perhaps a case in point here. While his framework for explaining “bad beliefs” is symmetrical—he explains them using mechanisms of general belief formation, not special mechanisms for forming bad beliefs—his examples have a clear political skew (climate sceptics, anti-vaxxers, creationists, and so on).

This suggests a further requirement: the political epistemologist needs a varied diet of examples. They should be as interested in understanding how mainstream liberals form beliefs as in how conservatives, socialists, populists, or any other political tendency you can think of, form them. One reason why I like Kahan’s work—despite some of the controversy surrounding it, including replication issues—is that he applies his cultural cognition framework to beliefs with which the reader might well agree as well as to beliefs that many of his readers are liable to view as false, perhaps dangerously so. This further requirement is a good way of cashing out what the reflexivity tenet—which I haven’t discussed at all—means in practice. A reflexive political epistemology will be interested in a wide range of political attitudes, beliefs and practices, including those attitudes, beliefs and practices that many political epistemologists endorse and engage in.

Since then, things have changed in analytic philosophy. We have gone through what some call the “applied turn” and others call the “social turn”. For its advocates, these turns demonstrate the practical value of philosophy, which is meant to lie in its critical potential; the tools of analytic philosophy can be used to understand and ameliorate an array of burning injustices. For its critics, these turns are a kind of cheapening of philosophy; at best, you get progressive politics with analytic precision. At worst, you get a subordination of philosophy to the imperatives of contemporary left progressivism. Underlying all this, as much else, are crude political oppositions—woke vs. anti-woke, progressive vs. conservative, left vs. right (or “centrist”).

I wrote recently about the destructive dynamics of politicisation. This is a case in point: those who are not fully on board with the progressive politics that are often viewed as driving the applied turn are left with the choice between ignoring it (hard to do when jumping on board might help you get a job) and fighting against it. The result is a pitched political battle—or a thin veneer of philosophy covering a pitched political battle—which is liable only to generate more heat (plus a few books). The closest I will ever come to writing about the “trans debate” is to note that this is the perfect example of what I have in mind.

Anyway, I think that this is the wrong way to think about the applied turn. You can cite examples of work in the applied turn that fits with this narrative; take, for example, Kate Manne’s celebrated work on misogyny or Sally Haslanger’s impressive body of work in feminist epistemology, metaphysics and philosophy of language (summarised in this book).

But you can also cite examples of work that, on any principled definition, belongs within the applied turn but either simply refuses a straightforward political reading, or advances a political cause that in some way or other departs from the shibboleths of contemporary left progressivism. Good examples include this book by Justin Tosi and Brandon Warmke, Maya Goldenberg’s discussion of vaccine hesitancy, Jason Brennan’s argument against democracy, Elizabeth Barnes’ book on disability, Hrishikesh Joshi’s work on Mill, liberalism and the problems with contemporary politics, C. Thi Nguyen's influential work on echo chambers and epistemic bubbles, or practically everything Dan Williams writes.

What, then, is the applied turn? Is it just a way of dressing up progressive politics in the guise of analytic philosophy? And is it the same thing as the social turn?

The Applied Turn: Top-Down vs. Bottom-Up

An initial problem with the idea that the applied turn is something new, a departure from 20th century analytic philosophy, is that the branch of ethics called applied ethics is usually tied to some very influential work in the 1970s, like Peter Singer’s Animal Liberation and “Famine, Affluence and Morality”. Since then, applied ethics has gone from strength to strength, as witnessed in the growth of bioethics, or the incredible impact of effective altruism.

The defining feature of this sort of applied ethics is that it takes established theoretical frameworks—moral theories like utilitarianism, political theories like libertarianism, or contractualist accounts of rights—and applies them to concrete moral and political questions. The aim is to derive practical conclusions—we should not eat meat or consume animal products, abortion is morally permissible—via a rigorous application of these theoretical frameworks. This can take a crude form, as when someone argues for a conclusion simply on the basis of utilitarian premises. Or it can take more complex forms, as it typically does in, for example, Singer’s work Still, the “direction of travel” is very much from high-level theory to concrete problem. Philosophical theories, sensibly applied, establish practical conclusions—they tell us what to do.

This suggests that the applied turn must mean something other than the application of theoretical frameworks to concrete questions. When you look at standard examples of the applied turn, like Miranda Fricker’s book Epistemic Injustice, you find an approach that seems to go in a different direction. Epistemic Injustice starts with some paradigmatic examples of epistemic injustice—Tom Robinson’s courtroom testimony in To Kill a Mockingbird, Marge’s testimony about Tom Ripley in The Talented Mr Ripley, the concept of sexual harassment—and then uses various theoretical tools to analyse these examples, which leads to a more general account of epistemic injustice. The direction of travel here is from concrete examples to some kind of theory. The theory is typically not at the general level of a moral theory like utilitarianism; it is a theory of a particular phenomenon, or a related bundle of phenomena. It aspires for a kind of generality, but it is more limited than the kind of extreme universality that is typical of more traditional work in philosophy.

While Fricker makes some use of conceptual tools from psychology, it is fair to say that most of her theoretical apparatus is drawn from philosophy, and in particular from 20th century epistemology (especially virtue epistemology) and some bits of philosophy of language. In this respect Epistemic Injustice is not the best example of the kind of methodological promiscuity that drives the applied turn. When we look at other examples, we see that one of the defining characteristics of the applied turn is an openness to drawing on machinery from a wide range of disciplines.

If we just stick to the examples I have already mentioned, we find those working within the applied turn drawing on evolutionary biology and psychology (Williams), cognitive science (Williams), disability studies (Barnes), economics (Brennan, Joshi, Williams), moral psychology (Tosi and Warmke), political science (Brennan, Joshi), science and technology studies (Goldenberg), sociology (Goldenberg) and social psychology (Joshi, Nguyen, Williams). If we broaden out and look at the applied turn more generally we can add several other disciplines: anthropology, linguistics, legal theory, history, and many more besides. Most importantly, the function of these other disciplines is not simply to provide a “check” against the claims made by philosophical theory; they provide the building blocks from which a theory of some phenomenon—vaccine hesitancy, as in Goldenberg’s book, misinformation, as in some of Williams’ work—is constructed.

This gives us a workable definition of the applied turn. It refers to an approach in philosophy that starts from the phenomena rather than some theory. That is, rather than deriving conclusions by applying a preexisting framework, it builds up from concrete social, political, or moral phenomena. Importantly, unlike the case- and intuition-driven approach that was common throughout 20th century analytic philosophy (and, if we’re honest, is still common now), the phenomena are richly described, often in terms taken from other disciplines. We start with the social phenomenon of grandstanding or vaccine hesitancy, which is presented in its complexity, rather than in the form of simple vignettes designed to elicit particular intuitions.

Another difference is that these phenomena are analysed and interpreted using whatever conceptual tools are useful, which means the theoretical work is often done by tools taken from economics, game theory, psychology or sociology, rather than analytic philosophy. The result is theory that is answerable to the phenomena it started from and that provides a kind of understanding and insight that is not possible just with the tools of analytic philosophy. It is hard to see how you could gain much of an understanding of something like grandstanding, misinformation or vaccine hesitancy armed only with these tools.

Politics and the Applied Turn

One common view—a view voiced in this piece—is that the applied turn is naturally allied with a particular political agenda. This agenda is, roughly put, that of contemporary left progressivism. A little less roughly, the agenda is the critique of existing injustices, with a view to ameliorating those injustices. Thus applied work in epistemology critiques epistemic practices that produce injustice, with a view to making those practices less unjust, while applied work in philosophy of language focuses on critiquing linguistic practices.

The value of this kind of explicitly political work depends heavily on two things: (i) what is being critiqued and (ii) the quality of the critique. Some of this work is, in my view, of a high quality; Charles Mills would be a good example of how to do this well. It is fair to say that not everything produced in this tradition is as good as The Racial Contract. But it is a very bad idea to judge the quality of a field by its worst—or even its average—outputs. The average paper in analytic philosophy is tightly argued, easy enough to read if you know the jargon, and makes, at most, an infinitesimal contribution to the sum of human knowledge. It may well be that the average paper in the explicitly political branch of the applied turn makes, at most, an infinitesimal contribution to eradicating any of the burning injustices that animate it. I fail to see why we should conclude that the average paper on epistemic injustice is worse than the average paper on the value of knowledge or modal epistemology. It may be that it is easier to write, but then it’s easier to write a paper on the value of knowledge than a paper on higher-order modal logic.

As I have been urging, though, not all work in the applied turn shares this particular political orientation. It may advance a different political orientation, like Jason Brennan’s case against democracy, or Elizabeth Barnes work on disability. Or it may resist a straightforward political reading, like Dan Williams’ work on misinformation.

Setting aside the political valence of work within the applied turn, we can draw a distinction between two ways in which applied work might have a political valence. The first way is when the politics is at the front and centre. A set of political commitments play a central role in the choice of phenomena and in setting criteria for a successful investigation of the phenomena. This might mean, for example, starting from a commitment to justice, to emancipation or liberation, to taking the perspectives of marginalised groups seriously. These commitments will then determine what you look at (presumably you will go looking for injustices) and what it would mean to find a productive lens through which to look at those injustices (you will want to say something about how they could be ameliorated). On this approach, politics is what motivates and orients the inquiry. This need not be viewed as subverting truth to political considerations. The political commitments may be viewed as true, so starting from them is a matter of advancing the truth, not ignoring it, or discarding it when it stops being politically expedient. Some of the examples of the applied turn I have mentioned clearly fall into this category: Fricker, Haslanger, Manne, Mills.

The second way is when an attempt is made to separate phenomena and an analysis of them from one’s political commitments. We need to be careful here; it is certainly true that it is difficult to achieve this sort of separation, and I am willing to agree with those who argue that, in a certain sense, a complete separation is impossible. But this fact—if it is a fact—clearly does not mean that there is no difference between an approach where one’s politics is front and centre, as a constitutive part of what one is trying to do, as in, for example, feminist epistemology, and an approach where one’s politics are in the background, and exert their influence in more subtle ways.

There are, for instance, presumably all sorts of political reasons why someone might be interested in conspiracy theories, misinformation or vaccine hesitancy. But there is a difference between engaging in an investigation of these phenomena while trying to keep one’s political commitments at arms distance and engaging in the investigation as, for example, an evangelist for universal vaccination, or as a card-carrying conspiracy theorist. Other of the examples of the applied turn I have mentioned clearly fall into this, second category: Goldenberg, Joshi, Nguyen, Tosi and Warmke, Williams.

This difference—between the social turn driven by a set of political commitments and the social turn with any political commitments kept at arms length—can perhaps be operationalised in the following way. A good indication that someone is engaging in the social turn with some political commitments in mind is when they draw on theoretical frameworks that have a clear political valence: feminist theory, for instance, but also political views that do double-duty as theoretical frameworks, like conservatism, liberalism or libertarianism. On the other hand, the sort of theoretical promiscuity I have highlighted as a defining feature of much of the social turn is an indication that someone is trying to keep their political commitments at more of a distance. This can be true even if you think that, for example, evolutionary psychology has an implicit politics; there is a difference between having an implicit politics and something being an explicit and elaborate political framework.

The Applied Turn vs. The Social Turn

The applied turn represents a shift in method and direction of theoretical travel—we start with concrete phenomena rather than principles and theories, where those phenomena are understood and investigated using a wide array of conceptual tools, including many tools not typically used by analytic philosophers. The driving principle is thus a combination of methodological pluralism and a commitment to “bottom-up” theory.

What I have called the applied turn is often—perhaps more often—referred to as the social turn. You can, within reason, use labels in whatever way you like, and there is no reason why you couldn’t refer to this shift in method and direction of theoretical travel as the social turn. But, as I’m using the labels, there is a big difference between the applied turn and the social turn. The social turn represents a shift in subject matter and theoretical commitments rather than a shift in method or direction.

In its most general sense, the social turn refers to the recognition that many of the things studied by philosophers—knowledge, language, agency, maybe value—are social phenomena, and so any theory which ignores this fact is simply mistaken. Thus, social epistemology refers to a movement which highlighted the ways in which social interactions can produce (as in testimony) or threaten knowledge (as in, for example, disagreement). It therefore represented a shift in subject matter (epistemologists now discuss testimony as well as perception) and theoretical commitments. This shift may be minimal, as when the theorist insists that theories of knowledge must accommodate testimonial knowledge, or maximal, as when the theorist insists that social epistemology requires a radical orientation of epistemological theory.

The applied turn is different from the social turn because taking the social turn does not necessitate taking the applied turn: you can recognise that agency or knowledge are social phenomena while continuing to study them in the way that philosophers did throughout the 20th century—by starting with toy cases and constructing philosophical theories around them with minimal insight from other disciplines and without any attempt to match those toy cases to concrete phenomena which might be better understood with the aid of other disciplines and the theoretical frameworks they provide.

It may also be that taking the applied turn does not necessitate taking the social turn. This is one way of understanding some naturalistic approaches in epistemology and philosophy of mind which draw on empirical sciences that tend to study individual humans in isolation from their social environment, like cognitive psychology or neuroscience. One could, for example, develop an applied epistemology of individual reasoning—how various cognitive biases distort or possibly enhance the beliefs of individuals without any interest in the social structures in which those individuals are located.

That said, most of the good examples of the applied turn that I have highlighted combine it with the social turn. The result is an approach that departs from much of 20th century philosophy in several respects: social rather than individualistic in theoretical commitments, promiscuous when it comes to method, and driven by concrete phenomena rather than puzzles that arise within existing philosophical theories and frameworks.

Why the Applied Turn?

The applied turn is a methodological shift—bottom-up theory allied with theoretical promiscuity—rather than left progressivism cloaked in the language of analytic philosophy. This is consistent with the fact that some work in the applied turn could aptly be described as left progressivism cloaked in the language of analytic philosophy. The problem is that there is also a lot of work in the applied turn that could not be described in this way, whether because it has a different political valence (think Brennan’s argument against democracy) or is at least trying to preserve a degree of political neutrality.

But why should we take it? It is tempting to fall back on answers that range from the self-interested to the downright cynical: it will appeal to students (true in my experience), help us get funding (probably also true), or at least make it easier to strike up a conversation with other academics. I’ve used these sorts of answers myself on occasion. But there is also a more principled argument to be made.

There’s a lot to be said for abstract philosophical theorising; at its best, it can be theoretically elegant, and we can learn things from it. But, occasionally, philosophers start to wonder what the connection is between their abstract theories and the real world. They ask questions like "why do people believe obvious falsehoods?", "what is so important about democracy?", "what is the point of moral talk", "why do some people's testimony get systematically ignored?”. It is tempting to think that the way to answer these questions is to take your off-the-shelf theory—your theory of rationality, your theory of testimony—and use it to answer them.

The value of the applied turn is that it demonstrates that this is a mistake. If you want to understand misinformation in the 21st century, you need to start with the ecosystems in which it exists and the psychological profiles that seem to predict acceptance of it. You also need to consider the long history of misinformation. Once you have done these things, the prospects of an illuminating general theory of misinformation might start to look a bit bleak. This is a good thing: the world is more complex than is dreamt of by any elegant philosophical theory.

The same is true across the range of phenomena that the applied turn has taken seriously. You cannot understand grandstanding by starting from a theory of assertion or moral discourse, or vaccine hesitancy by starting from a theory of testimony or trust, or the epistemic dynamics of echo chambers by starting from a theory of rational belief revision. In each case, the phenomenon itself is what is worth engaging with, and it is of sufficient complexity as to require a range of theoretical machinery to understand.

The point is not that we should replace traditional philosophy with the applied turn. The point is that, if we want to understand real-world phenomena—our propensity to believe pleasing falsehoods, the crises of democracy—we need to do it properly. This leaves me in the unusual—for me—position of advocating a policy of extremes. You want to keep your traditional philosophy as traditional as possible and your applied philosophy as applied as possible. The problem is when you try to have it both ways: the elegance of traditional philosophy directed at real-world problems. If the result is elegant, it won’t help you understand the problem. If it helps you understand the problem, it probably won’t be very elegant.

I want to answer two questions:

1. What is politicisation?

2. What, if anything, is wrong with politicisation?

It is important to not start from the assumption that politicisation is a bad thing. First, whether it is a bad thing depends on your political ideology. On the one hand, some ideologies are suspicious of it, for example, free-market conservatives, who want to leave things to the market. There is also a technocratic kind of liberalism that is suspicious of politicisation, on the grounds that it means handing issues over to the public that might be better answered by the application of specialist knowledge. On the other hand, there is a more democratic kind of liberalism that is in favour of politicisation, providing it takes the right kind of form, and populists, whether left or right, tend to decry depoliticisation—the shifting of important questions out of the political arena and into the arena of technocratic governance.

Second, politicisation is not always a bad thing. Imagine a society where slavery is widely practiced and it is not regarded as a matter for public political debate. In this society, one would want the question of slavery to become politicised. In any society where political arrangements could do with real improvement (any human society that has ever existed), it is important that there are avenues available to shift problematic social arrangements into the domain of political concern.

If we don’t start from the assumption that politicisation is a bad thing, what should we do? We want a minimal definition of politicisation that is not geared towards the context of democratic politics but applies in a wider range of political contexts. But, because I don’t think it is possible—or desirable—to say what is wrong with politicisation without assuming any sort of political ideology, my approach will be to try and make explicit the political ideology underlying different attempts to identify conditions under which politicisation is a problem. Specifically, I will contrast two different ways of thinking about the problem with politicisation, corresponding to two different underlying conceptions of politics and the political.

The first, which is broadly liberal, is represented by Robert Talisse’s recent book Overdoing Democracy, which I have written about previously. The second, which is broadly republican, highlights some things that Talisse misses. I will develop this conception by drawing on Albert Hirschman’s influential discussion of exit and voice in his book Exit, Voice and Loyalty, which, I argue, gives us an account of the conditions under which politicisation might enhance or curtail political agency.

What is Politicisation?

There is a nice paper by Rich Eva where he goes over all the definitions of politicisation in the literature. He finds all these definitions wanting, and provides his own definition. It doesn’t seem to be available open access (there is a paywalled version here) so you’ll have to trust my brief overview of it.

Eva rejects several definitions of politicisation on the grounds that they give the wrong verdicts about cases. He wants a definition that includes cases that lie outside of institutional politics; for example, a football player using a political slogan, or an actor making a political speech at an awards ceremony. He thinks that politicisation is closely connected with partisanship. Thus his initial definition: “to politicize is to use a nonpartisan good as a means to a partisan end.” He expands:

To politicize, one must use a nonpartisan good as a means to a partisan end, where a partisan end is defined as a goal geared toward the exercise, attainment, or maintenance of governing power by a contested political party, candidate, or political ideology.

A nonpartisan good is a good not essentially tied to partisan politics. Examples of nonpartisan goods include sports, religious practices and cultural practices, but can also include state and government agencies. A government agency is not inherently partisan in Eva’s sense because you would still need some government agencies in a state without partisanship, like a one-party authoritarian state.

The problem with this definition is that you can have politicisation outside the sphere of partisan politics—or even without any sort of open conflict or dispute. Some of the most extreme forms of politicisation occur precisely where democracy has been suppressed.

A particularly striking example comes from Václav Havel’s famous essay The Power of the Powerless and his example of a greengrocer, whose everyday actions, like hanging a poster saying, “Workers of the world unite!”, are laden with political meaning and significance despite the absence of any kind of partisan politics. In this context, there is politicisation without any partisanship—or indeed any dispute. The aim is the opposite; it is to produce complete conformity in behaviour.

One reason why Eva can’t accommodate this sort of case is that he is primarily interested in how individual agents can act so as to politicise things (issues, institutions, practices). Havel’s greengrocer is not responsible for the fact that his hanging the poster is laden with political significance; you are to assume that, if he gives the matter any thought, he resents it. The point is rather that he inhabits an environment where actions like this are already imbued with political meaning—putting up the poster signals that you are a compliant citizen, whereas not putting it up singles you out as a dissenter. This suggests that what we need is something broader, not so focused on individual acts.

My own account of politicisation is modelled on Oliver Traldi’s account of the politicisation of beliefs in his book Political Beliefs. He tells us that:

A belief that is disputed can be politicized by gaining the right kind of connection to politics; a belief that has the right kind of connection to politics can be politicized by becoming a matter of dispute; and a belief that meets neither of these conditions can be politicized by gaining both.

Where saying that:

a belief is political requires that it be in dispute and that it have the right kind of connection to politics; two ways of having the right kind of connection to politics include figuring in the right way in practical reasoning about specifically political action and being an important factor in an agent’s inclusion or exclusion from at least one political aggregate or group (p. 19).

This is a disjunctive account: for a belief to be politicised, it must both have become a matter of dispute and have the “right kind of connection to politics”. It is not obvious why we need the first disjunct. Take the belief that immigration needs to be reduced. So long as this figures in practical reasoning about political action (e.g. it leads people to vote for a party that promises to do this) or functions as a badge of inclusion or exclusion in a political group (e.g. you need to profess this belief to be included), it doesn’t seem to matter how much dispute there is about immigration.

Amending Traldi’s definition, so that a belief is politicised so long as it has the right kind of connection to politics, whether or not it is currently in dispute, has the additional benefit that it can then be extended to the politicisation of things that are not easily understood as potential matters of dispute. This includes things like activities (e.g. the activity of hanging a poster in a shop window). While activities cannot in any straightforward sense be matters of dispute, and are the upshots of rather than premises in practical reasoning, they can be important factors in group inclusion and exclusion—you can be excluded from a political group for not performing the activity, and your inclusion can be conditional on performing it.

I therefore propose these definitions:

Politicisation is the process by which something (e.g. an activity, belief, issue, institution) acquires the right kind of connection to politics (or reacquires it after having lost it).

Something (e.g. an activity, belief, institution, issue) is politicised when it stands in the right kind of connection to politics.

The first is a definition of the process of politicisation—something is politicised through acquiring the right kind of connection to politics. The second is a definition of the state of being politicised. These definitions seem to accommodate all the examples of politicisation we have discussed. Most importantly, on these definitions the greengrocer’s act of hanging the poster in his shop window is politicised because it has the right kind of connection to politics—it is vital to the greengrocer not being marked out as some sort of dissenter.

The obvious problem, though, is that these definitions verge on the trivial; it is not surprising to learn that there is a close connection between politicisation and having the “right kind of connection to politics”. But, as I said at the beginning, we want this kind of triviality. We want to avoid defining politicisation in such a way that it assumes any particular conception of the political.

Talisse on Political Saturation

This issue—implicitly assuming a particular conception of politics—also arises when it comes to thinking about what might be wrong with politicisation. This point is exemplified by Robert Talisse’s recent book, Overdoing Democracy, which illustrates how an underlying political ideology can shape an account of the problem with politicisation. (For a longer version of this argument, see my earlier post).

In Overdoing Democracy Talisse tries to explain the political crisis that many think we (in much of the Global North) are experiencing. The nature of this crisis isn’t clearly specified, but the crisis of knowledge, along with the broader politicisation of institutions, is an important part of it, as are familiar phenomena like polarisation and rising distrust, even revulsion, in politicians and our political opponents. Talisse thinks the cause of these problems is that we are overdoing democratic politics. Rather than viewing politics as something we must do to create the necessary background conditions for living meaningful lives, politics has become the focus.

One of Talisse’s central concepts is “political saturation”. Political saturation is “the saturation of social life with activities and projects that are overtly organized around the categories and divisions of current politics.” Social spaces (dinner tables, coffee shops, lecture theatres) and the institutions and organisations that provide these spaces (like universities) have become venues in which we perform our political identities. Talisse thinks that this is a problem because, in these politically saturated spaces, we are either just performing to like-minded partisans or antagonising rather than engaging our political opponents. The result is that our political activities are either pointless (when we’re preaching to the choir) or actively destructive (when we’re antagonising our opponents) of the social fabric that holds democracy together.

Talisse’s conclusion is that, if we want to improve things, we need to do something else—something that isn’t politics: “More and better politics cannot be the solution to the problem … because politics is the problem“ (Talisse 2019, 7). Talisse recommends that we look for things to do together that have nothing to do with politics, though he admits it is hard to say what these things are, because political saturation tends to colonise all aspects of our lives.

While I agree that political saturation is a serious problem, I am sceptical that this is due to us “overdoing politics”. Talisse’s argument for this claim is based on his underlying conception of democratic politics, which is deliberative:

The central deliberativist thought is that, in a democracy, collective decisions derive their authority from the fact that, prior to voting, each citizen was able to engage in processes whereby he or she could rationally persuade others to adopt his or her favored view by defending it with reasons and offering reasons opposing competing views. According to the deliberativist, then, the democratic ideal has as its core an idea of collective reasoning.

This is a conception of democratic politics that, at least in principle, doesn’t recognise many limits. Any collective decision requires deliberation, which, because it is rare to achieve genuine consensus, will often take a long time and get quite heated. The result is that the entire social sphere is within the reach of democratic politics and we end up “doing politics”—having deliberations, often to no real effect—all the time. Given this conception of democratic politics, Talisse’s suggestion that we do less of it makes sense.

There are however some problems with this way of understanding politicisation, political saturation and the problems with democratic politics that are driving it. The first is just a version of the point made earlier: political saturation, like politicisation in general, is possible outside the context of democratic politics and deliberation. Indeed, we can have extreme forms of political saturation under conditions where there is little or no public political deliberation.

Consider again Havel’s greengrocer. Havel’s argument, in brief, is that what he calls a post-totalitarian state (like Czechoslovakia at the time he was writing) represents an extreme form of political saturation, where mundane activities and decisions are imbued with political meaning. For Havel, this is how a post-totalitarian society maintains itself: it produces conformity in behaviour and practice that suppresses dissent and creates the illusion of widespread public support for the regime. Political saturation here has nothing to do with an excess of deliberation or politics. If anything, it is a consequence of the complete absence of political activity as it would be understood in a “healthier” society. The only way in which someone can “contribute” to politics in this context is by performing the role (compliant citizen of a communist state) required for survival in this context, with the result that their compliance is far from freely chosen.

This suggests that the connection between politicisation, political saturation and democratic deliberation may be weaker than Talisse seems to imply.

Second, Talisse’s conception of democratic politics—on which, put crudely, it involves talking about politics—contrasts with conceptions that foreground more direct forms of political action, like voting and canvassing for political candidates/parties, standing for elected office, participating in local government, community organising, participating in trade unions or other forms of collective bargaining, participating in protests or other forms of civil disobedience, and lobbying, petitioning or otherwise exerting direct pressure on decision-makers. This is essentially the difference between participating in politics as a spectator and participating in politics as an actor.1

If we focus instead on participating as an actor, this gives us a different way of thinking about political saturation, on which it is the result of a deficit of political action. Specifically, it is the result of two factors. The first is a lack of avenues for genuine political action. There is an American version of this story, on which the story of American democracy is—with a few interludes—a story of elite capture and managed democracy, and a more global version, which highlights globalisation, neoliberalism, and the resulting democratic deficits at a national level.2 The second factor is a technological and social landscape that provides plentiful opportunities for pale surrogates for genuine political action, like arguing with strangers online about politics. The result is that the performance of political identity comes to substitute for genuine political agency, and that substitution occurs because avenues for genuine political agency have been closed off and replaced with surrogates that have the “look” of politics without any of the substance.

The difference between these two ways of thinking about political saturation—as an excess of politics or as a deficit of genuine political agency—maps directly onto two different conceptions of politics and politicisation. On Talisse’s deliberative conception, the central political value is legitimacy, and politics is fundamentally about reason-giving and public justification; political systems, and the decisions they produce, are legitimate and have authority over us only insofar as they can be justified through public reason. On this conception, to politicise something—to give it the right kind of connection to politics—is to bring it into the space of public deliberation and make it subject to demands for public justification. What this means depends on what is being politicised. To politicise an issue is to make the various positions one might take on it subject to this demand for public justification. To politicise something like an institution or a practice will mean something else; it might, for example, mean public deliberation about the role of the institution or practice, or about the need for it. Either way, politicisation always means bringing public deliberation to bear on whatever it is that is being politicised.

On our alternative conception, which can aptly be described as republican, the central political value is non-domination—the absence of arbitrary power over others—and politics is fundamentally about the ways in which power can be checked, contested, and held to account. What matters on this view is not so much the ways in which decisions and institutions are justified as that those who are affected are able to contest and challenge them. On this conception, to politicise something is to contest it through organised action. Again, what this means depends on what is being politicised. It is perhaps easier here to see what it might mean to politicise something like an institution—it would be to contest and challenge the power it holds over us, or the decisions it makes and enforces. To politicise an issue will mean something else; for example, it might mean contesting the terms in which the issue is discussed, or drawing attention to who benefits from it being the subject of public deliberation. Either way, politicisation is always tied to some form of direct political action.

To clarify the differences between these conceptions, we can return to Talisse’s discussion of political saturation. On the deliberative conception, political saturation is a process of increasing politicisation; as social spaces become arenas for political discussion and debate, they become politicised. As more social spaces become arenas for political discussion, the more politicised a society becomes. At an extreme, we can imagine a society where every social space—including social spaces within the private sphere, like a dinner table at a holiday celebration—is an arena for political discussion. Such a society would be maximally politicised.

Things look a little different on the republican conception. Whether these social spaces are politicised depends not so much on whether they are arenas for discussing politics as on whether these spaces facilitate direct forms of political action. A space that has no connection with direct political action is not in any real sense politicised. Political saturation is therefore only a process of increasing politicisation if it is accompanied by an increase in the possibilities for direct political action. These two things may go together, as, for example, in a situation where a workplace becomes heavily politicised and this leads to a strike or some other form of direct action. But they may not, as in a situation where someone spends their evenings arguing about politics online, but those arguments achieve nothing other than a general increase in anger and aggravation.

These differences matter because they are reflected in different stances on the question of what, if anything, is wrong with politicisation. While both the deliberative and republican conceptions provide ways of understanding why politicisation might be necessary, beneficial, or counter-productive, they do so in very different ways. On the deliberative conception, politicisation is required for collective decision making, and it is necessary because in politics we need to make a lot of collective decisions. Problems arise when public deliberation begins to degrade our capacity for collective decision making, as Talisse thinks happens under conditions of political saturation. Talisse is worried that, when we take it too far, public deliberation ends up eroding and threatening the social trust required for deliberation to be productive. His solution to this problem is the natural one, given his underlying understanding of politics: we need less deliberation, or at least we need to restrict deliberation to contexts and situations where it can be productive.

On the republican conception, on the other hand, politicisation is required for holding power to account, and it is necessary because—at least if we want to avoid various forms of domination—citizens need means of holding power to account. Views on when politicisation becomes a problem will depend on how much sympathy one has for the underlying republican conception of politics. Presumably everyone is concerned with some forms of domination, so will agree that, for example, when a practice or institution has become completely unresponsive to any public concerns and is regularly interfering in the lives of citizens, there is a need to politicise it—to subject it to some sort of challenge (consider our earlier example of slavery). Someone who is not particularly sympathetic to republicanism will however hold that politicisation often goes “too far”: sometimes, it is good that ordinary citizens have no means of contesting political decisions (this is the perspective of the free-market conservative or technocratic liberal).

Rather than arguing for one of these conceptions over the other I want to suggest something more pluralistic. A pluralistic approach to the question of politicisation—that is, one not tightly wedded to either a deliberative or republican conception of politics—should focus on what is valuable in these different conceptions. The deliberative conception highlights the need for politicisation in contexts where the basic question is one of legitimacy. But this conception is not suited to thinking about some of the problems that arise in deliberations that are nominally supposed to secure that legitimacy but have no prospect of doing so. When there is a clear democratic deficit and public deliberations—like political discussions on social media—lack any connection to political decision-making the problem is not “too much” deliberation but rather that the deliberation doesn’t do anything. It certainly isn’t an effective means of contesting or challenging power; it is a side show. This suggests that the republican conception of politicisation might be particularly useful in contexts of dysfunction, where it is clear to many that existing arrangements are not working and that powerful actors are not being held to account.

Politicisation, Exit, Voice

On the republican conception, politicisation may be required to challenge and contest power. By, for example, politicising an institution like a central bank, you can open it up to challenge and contestation. But simply contesting things doesn’t necessarily achieve anything. The key question is: when will a means of contestation be effective, and when might it become counter-productive?

To discuss these issues I want to use some ideas from Hirschman’s Exit, Voice, and Loyalty. Hirschman’s central idea is that we can understand responses to decline in businesses, institutions, organisations and states in terms of the differential costs and benefits of two kinds of response: exit (not using a company’s products, quitting an organisation) and voice (making dissatisfaction known, protesting, speaking out). While you can have exit in politics, and voice in business, Hirschman sees exit as closely tied to consumer behaviour and voice as closely tied to political behaviour. It is perhaps better to say that the paradigmatic way for a consumer to express dissatisfaction with a product (not buying it) is a form of exit, whereas some paradigmatic forms of political behaviour (like arguing for a different policy, or advocating for a change in leader) are forms of voice. But it isn’t hard to think of political behaviours that qualify as a form of exit: leaving a political organisation, protest, not voting, spoiling a ballot paper, striking, or not using services provided by the state (e.g. opting out of a state education or healthcare system).

A virtue of Hirschman’s analysis is that he details the complicated dynamics between exit and voice. I want to highlight two of these dynamics. The first is that, even if a lot of people take the exit option, this need not produce any change or improvement. For example, if a lot of people leave an organisation because they are dissatisfied, then the organisation may end up staffed by those who were still relatively satisfied (so they won’t exercise voice), or those left behind may not have a meaningful exit option, with the result that the organisation will not be responsive to their exercise of voice. Either way, this means that the organisation will not change. Whether this is a problem, of course, depends on the organisation, the nature of the dissatisfactions that led people to leave it, and whether it serves a valuable function. If the dynamic plays out at the level of a whole country, as when many opponents of a regime emigrate, the results can be disastrous, at least for those left behind.

The second dynamic concerns the impact of voice. Hirschman says that voice only effects change in an organisation under two conditions: the organisation must be sufficiently responsive that voice can produce change, and exit must be a meaningful alternative to give voice leverage. When an institution is unresponsive, or exit is unavailable, voice has little effect. Imagine, for example, an organisation that simply does not want to listen to the concerns of its members, in part because those members have nowhere else to go (perhaps this is the only organisation that could represent their interests). No matter how vocally the members give voice to their dissatisfaction, it is unlikely that much will change. Things would be very different if there were a rival organisation that also had a claim to represent the interests of the group members. This would give them an exit opportunity, and therefore more leverage.

We can apply this framework to politicisation in two ways. The first is as a way of thinking about when politicisation is beneficial—when it reverses decline within an organisation or institution. On the republican conception, politicisation is the process of bringing something into the sphere of contestation. When contestation is beneficial, exit and voice combine to produce improvements. Consider the thinking behind strikes and other forms of withdrawal of labour. The idea is that the threat of (temporary) exit in the form of a strike will make the employer more responsive to workers’ concerns. Whether this works in practice depends on the situation; some strikes are hopeless because the employer won’t, or simply can’t, respond to the workers’ concerns. Or, to use an example with a different political valence, consider a political campaign against tax increases that leverages the ability of wealthy taxpayers to leave the tax jurisdiction. The idea again is that the threat of exit will make the government more inclined to listen to campaigners’ arguments against tax increases. In both cases, politicisation produces change because of the combination of exit and voice. Of course, whether that result is viewed as an improvement depends on your political ideology—this goes back to the point that it is hard to think about politicisation without assuming a background ideology.

The second, and for our purposes more important, application is as a way of thinking about when politicisation becomes problematic. Even if we set aside the ways in which, depending on your underlying political ideology, particular kinds of politicisation will be problematic, we can identify some ways in which politicisation can be destructive of institutions. Put briefly, the politicisation of an institution becomes destructive when it reduces the possibilities of using exit and voice to remedy problems with the institution.

The basic idea is that, once an institution becomes politicised, and acquires some sort of political agenda, this has implications for the use of exit and voice to remedy any of the problems with it. To see how this works, consider someone who is unhappy with the institution in its current politicised state. On the one hand, if they are within the institution they will recognise that there are many costs to using the exit option. Exiting the institution—or staying with it and reverting to a kind of apathy that is functionally equivalent to exit—would likely have the result that the institution is increasingly captured by the new political agenda, as those who oppose it are far more likely to leave than those who support it or are neutral. It may be that those outside the institution can use the exit option to greater effect; if the continuation of the institution requires it providing some services to the public, then the public can make their dissatisfaction known by refraining from using those services. This though depends on the broader context, and in particular how reliant the institution is on the consumers of whatever services it provides. On the other hand, someone who is happy with the way in which the institution has become politicised is unlikely to use the exit option, whether they are inside the organisation or make use of some service it provides. This is not just because they are aligned with the political agenda now driving the institution; exit would weaken the very agenda they are trying to advance.

The result is that once politicisation gets going, it is hard to stop because it creates a spiralling dynamic. This can happen in a few different ways. The main one is that, despite the costs, enough of the critics of the institution leave, with the result that the members of the institution are, as a whole, more supportive of the political agenda than they were before. This is politicisation as homogeneity. It may however happen that the critics decide to stay within the institution and fight, with the result that they end up pursuing their own, different, political agenda. This is politicisation as polarisation. Either way, the result is more politicisation, whether of the homogenous or polarised variety.

This spiralling dynamic is driven by the fact that politicisation tends to structure the discursive environment in ways that reduce the effectiveness of voice. One way in which this happens—the way just outlined—is indirect: voice is most effective when there is a meaningful possibility of exit so, by increasing the costs of exit, politicisation reduces the effectiveness of voice. Another way in which it can happen is more direct: politicisation undermines the effectiveness of voice via self-censorship and selective uptake. In a politicised institutional environment, the costs of exercising dissenting voice can be extremely high; there is a risk of dissent being viewed as running counter to the political agenda of the institution, and so dissent can lead to reputational damage, or even exclusion. The result is that internal critics start to self-censor, and the institution becomes less responsive to criticism, even criticism that it would benefit the institution to take seriously (for a development of similar ideas see this excellent book). This is exacerbated by the fact that, when voice is exercised in a politicised institutional environment, it tends only to be heard by those who already agree with it, and it is often misinterpreted to fit with prior perceptions, with the result that the institution only hears the criticisms it wants to hear.

These two pathologies—the increase in cost of exit and the hollowing out of voice—give us a precise diagnosis of a way in which politicisation can become problematic. Politicisation is a problem when it simultaneously raises exit costs and undermines the effectiveness of voice. The result is that institutions become less responsive, even to legitimate criticism, and the people who interact with those institutions lack effective means of contesting or changing them.

The Crisis of Knowledge

The obvious place to apply this analysis is where we started — the crisis of knowledge and trust in scientific institutions. Any attempt to apply this analysis to an example is liable to prove controversial. I do however think that the analysis can be usefully applied to the politicisation of knowledge and knowledge institutions.

In a recent paper Hrishikesh Joshi argues that, while the university system exerts enormous power over public discourse, it is democratically unaccountable and in that sense democratically illegitimate. Joshi’s argument is based on the political homogeneity of the academy. I want to suggest that we think of the issue differently: the university system has too few mechanisms for translating public dissatisfaction into political action.

Exiting the university system is costly for a student, who is deprived of the opportunity to enhance their future earning potential, not to mention to enrich their life. While the ordinary citizen cannot exit the university system in this sense, they can refuse to recognise the authority of scientific knowledge and scientific institutions, which deprives them of valuable knowledge and understanding that they could otherwise gain. Because there are few good exit options, and few mechanisms for giving voice to dissatisfaction, the result is that the public is often unable to translate any dissatisfaction with the university system into political action. Indeed, the forms in which public dissatisfaction with the university system is often expressed—online vitriol, for example, or the embrace of “alternative experts”—are exactly what one would expect when there are few mechanisms by which ordinary citizens can turn their dissatisfactions into productive political action. In this respect, the situation is like that I highlighted in my discussion of political saturation; in the absence of channels for political action, there is a tendency to turn to antagonistic behaviours that serve to inflame rather than resolve tensions.

This analysis suggests what a “solution” to the crisis of knowledge and trust in knowledge institutions would need to do: it would need to create conditions for exit and voice to function effectively.

Citizen science has been proposed as one way to give ordinary citizens more of a role in the production of scientific knowledge. But, as Chloé de Canson has argued in an important recent paper, citizen science is often understood in a way that means the involvement of ordinary citizens is relatively limited. In what de Canson terms “participatory science”, laypeople are invited to participate in research projects that are directed by scientific institutions and the scientists who staff them. This means that citizen science often proceeds entirely under the purview of institutional science; participatory “citizen scientists” can offer their opinions but have no real decision-making power. This is, essentially, voice without leverage, which produces a situation where the impact of voice is minimal, as there are no mechanisms for forcing uptake.

What de Canson calls “extitutional science” is more promising because it goes beyond invited participation. De Canson gives an account of communities of HIV/AIDS and ME/long Covid patients who formed what I have called “research collectives”—communities outside mainstream scientific institutions with their own internal structures for assessing and producing knowledge (I now have an academic paper developing these ideas). Extitutional science differs from participatory science in that these research collectives set their own research agenda, which may involve contesting and challenging the methods and background assumptions of institutional science. These research collectives seek to force institutional uptake through what de Canson, following Sarah Schulman, calls the “inside/outside strategy”: one arm of the movement collaborates with institutional scientists, while the other exerts political pressure to force a response. That this sort of strategy is liable to be more effective than a more confrontational “outside strategy” follows from Hirschman’s analysis: the combination of collective voice and external pressure provides more leverage than collective voice alone.

My suggestion then is that extitutional science provides one mechanism by which ordinary citizens can contest the findings and methods of knowledge institutions. The challenge is to apply this suggestion consistently. It is not hard to see the value in the activities of HIV/AIDS patients in the 1980/90s, who made valuable contributions to medical understanding in a political climate that was, to say the least, hostile. But what are we to say about cases of extitutional science during the Covid pandemic that challenged the then-standard views about the efficacy of measures such as masks and school closures? It is important not to dismiss these cases simply on the grounds of disagreement with the claims made by these communities, or with the political agenda one might presume to lie behind those claims. Extitutional science is often driven by a political agenda. What we need to avoid is picking and choosing the “good” cases of extitutional science in a way that simply affirms and reinforces our existing views and political sympathies.

“Political epistemology” has recently emerged as an area of analytic epistemology. In just the past few years there has been a handbook, an edited volume, and two textbooks published. Countless articles, books and book chapters have appeared on topics including conspiracy theories, echo chambers, fake news, the epistemology of democracy, expertise and trust in experts, political disagreement and polarisation, and irrationality in politics. Epistemologists have written, in both the academic journals and popular press (do we still call it the popular press?) about recent political events, such as the Brexit referendum, the election of Donald Trump, the rise of a global far-right, and the Covid-19 pandemic.

This is clearly a trend, and it is natural—I would say healthy—to view trends with a degree of suspicion. Part of the reason for the emergence of political epistemology is a preoccupation with recent political events (you know the ones), and the role that conspiracy theories, echo chambers, fake news, propaganda, polarisation and the like have supposedly played in them, in the UK, US, and a few other predominantly English-speaking countries in the Global North. I’ve often worried that there is something a bit parochial about this, both in the focus on the present, and in the focus on a small number of, albeit disproportionately powerful, countries. There something faintly puzzling about treating these phenomena as surprising, or as standing in need of special explanation. Anyone with even a passing acquaintance with 20th century history—Nazi Germany, Stalinist Russia, Maoist China—already knows that people are capable of accepting, and acting on, almost any set of beliefs under the right conditions.

One of the consequences of this is that political events in other countries, or political events that don’t fit with a preferred narrative, tend to get ignored. There is a lot more in the way of professional rewards for work on some political contexts and certain political problems than others, to such an extent that an early career researcher is probably best advised not to work on political contexts and problems that the profession at large regards as insufficiently interesting, or as suggesting the kind of politics of which some academics might disapprove. Put simply, you can publish your philosophy book on the pathologies of contemporary US politics with the leading university presses, but you will have a harder time if your focus is China, Russia or—god help you—a small country that nobody knows anything about.

All that said, I also want to suggest that there need not be anything wrong with parochialism, whether in general or in political epistemology. Moreover, something approaching parochialism is positively valuable: political epistemologists should be interested in political problems in particular countries, as well as the local flavour that more universal problems take on in particular countries. The problem is not parochialism as such. One kind of problem is when political issues and problems specific to a particular country are treated as if they were universal, or when problems that are common to many countries are viewed exclusively through a lens that makes sense in one particular country (you can probably guess which one). Another kind of problem is when you have an academic reward system that gives more accolades to work on some topics than on others.

What is Political Epistemology?

Boundaries are usually blurry, whether between disciplines or sub-disciplines. But you can typically cite clear examples on either side of a divide. Work on epistemic conceptions of democracy counts as political epistemology; work on foundational issues in the epistemology of testimony doesn’t. Other examples are harder to classify.

One helpful way of defining political epistemology has been suggested by Michael Hannon and Elise Woodard. Political epistemology is concerned with both (a) epistemological dimensions of political issues and (b) political dimensions of epistemological issues. Under (a) we consider topics such as epistemic conceptions of democracy, polarisation, political ignorance, and dis- and misinformation. Under (b) we consider topics such as expert disagreement, trust and mistrust in experts, and political disagreement. This helps explain why political epistemology is so tricky to characterise: it typically overlaps with political philosophy (under a) or social epistemology (under b).

A distinctive feature of political epistemology is that it is deeply engaged with empirical disciplines: economics, media and communications, political psychology, political science, sociology, and others besides. This is, at least in my view, one of the best things about political epistemology. It is outward-looking, far more so than more traditional epistemology. It is of interest to non-philosophers, perhaps even non-academics. Political epistemologists don’t just exchange intuitions about bank cases; they engage with actual data about how people form beliefs, who they trust, and why.

But political epistemology is liable to inherit problems from the disciplines on which it draws, which are often parochial in many of the same ways in which political epistemology is parochial. Most of the empirical studies discussed by political epistemologists concern the US, and concern political issues that either are specific to the US or take on a particular shape there. This is partly a data problem and partly a language problem: a lot of the relevant data is collected in the US, and most of it is available in English. The result is a field that knows a great deal about how, for example, polarisation works among American voters and less about how it works anywhere else.1

Parochialism

The OED defines “parochial” as “relating or confined to a narrow area or region, as if within the borders of one’s own parish; limited or provincial in outlook or scope.” In this sense, parochialism is a matter of one’s interests being confined to the local, to where one lives or calls home.

I want to urge a distinction between two attitudes, one of which is far more problematic than the other. There is a difference between taking a special interest in the narrow area one happens to call home and thinking that one’s own region is more important or interesting than others. We might call this the difference between parochialism and myopia. Both parochialism and myopia involve a narrowing of horizons, but the parochial stance need not involve any dismissal of other horizons, whereas such dismissal is definitional of myopia. The parochial individual is interested in the local, in what is around them. The myopic individual thinks that all that matters is what is around them.

Myopia is far more problematic than parochialism. Parochialism can be entirely harmless, and sometimes even positively valuable. As John Tomaney puts it in a defence of parochialism, it “values the local, its culture and solidarities, as a moral starting point and locus of ecological concern and a site for the development of virtues including commitment, fidelity, civility and nurture.” Of course, one can value the local in ways that are problematic, and refusing to broaden one’s horizons can reveal that one is actually myopic rather than merely parochial. But the crucial point is that myopia is always problematic whereas parochialism need not be.

The most problematic kind of parochialism is the kind that isn’t conscious of itself as parochialism. There is a difference between being primarily interested in events within your own country while recognising that things are different elsewhere and those differences also matter, and being convinced that what you are interested in is what everyone should be interested in, that the political dramas unfolding around you are of paramount importance to the world. The first is parochialism. The second is myopia.

What would it mean to say that political epistemology is parochial in this sense? Political epistemology is parochial to the extent that it focuses on political events in particular countries, or assumes a framing of more universal events that is specific to those countries. This might mean focusing on what is happening in certain countries, on the political dynamics within them, or on aspects of their citizens’ psychology, and treating the results as if they had general application. The election of Trump, the Brexit referendum, the pandemic as experienced in the anglophone world: these are often discussed in isolation from anything else happening in the world, or in a way that foregrounds them over the broader social, political and historical processes that might be part of a proper explanation of them.

It seems clear that political epistemology is parochial, at least to some extent. In saying this I am not necessarily criticising work on any of these events, still less dismissing it. Work on Brexit or on Trump’s election can be perfectly good political epistemology. What would be worthy of criticism is a situation where one researcher can easily publish a philosophically informed analysis of the Brexit referendum, yet another researcher can find no takers for an equally philosophically informed analysis of, say, the epistemological dimensions of democratic backsliding in parts of Europe, or of the role of misinformation in Brazilian politics. This situation is not far from reality—indeed, this is the reality. It is easier to publish a discussion of misinformation and its role in the contemporary right-wing US political ecosystem—often thinly disguised as a more general discussion of the “problem of misinformation”—than a discussion of misinformation in any number of other political ecosystems.

The simplest explanation for this is not that political epistemologists are myopic. It is that most of them work or were trained in the UK or the US, and political events in those countries are particularly salient to them. This is parochialism, not myopia, and parochialism—as I have argued—need not be problematic in itself. But it can become a problem when combined with other things, or when everyone in a field is parochial in more or less the same ways. Mere parochialism at the level of individual researchers can become myopia at the level of the field.

A Method for Political Epistemology

So far I’ve just been taking some potshots at political epistemology. What do I think should be done differently? I think the right method for political epistemology involves a good deal of focus on narrow geographical areas and regions. The crucial point is that we need sufficient variety of local contexts across the field as a whole.

It is useful to draw a distinction between two parts of political epistemology. The first is concerned with foundational questions: what, if any, is the epistemic value of democracy? What is the appropriate stance to take towards experts? How should we respond to political disagreement? The second is concerned with how these questions play out in particular social and historical contexts, and with questions that only really arise when one tries to make sense of specific contexts: what drives acceptance of particular conspiracy theories in particular communities? Do new communications technologies change socio-epistemic dynamics in ways that differ across different political cultures?

Much like any distinction, this one can become blurry in practice. But it is an instance of a more general distinction between ideal and non-ideal epistemology. The ideal part is interested in foundational problems and basic ideals and values. The non-ideal part is interested in how those ideals and values might be realised in practice, the problems one faces in trying to realise them, and the epistemological problems that only become visible once one focuses on the real world with all its imperfections. The relationship between the two is complicated (I tend to think you can often profitably engage in one without engaging in the other) but there are also genuine connections.

Even at the level of ideal theory we can identify tensions between our basic values. I’m particularly interested in the tension between democratic ideals and what we might call the “technocratic ideal”: policy-making informed by the best available scientific evidence. Put simply: giving more decision-making power to technocrats means taking it away from ordinary citizens. While this can be massaged by adopting various means for getting those citizens “on-board”—manufacturing consensus on the policies proposed by technocrats—the basic tension remains.

One can say quite a lot about this tension at a relatively abstract level. But at some point you might want to look at the particular contexts in which it manifests, which is where the non-ideal part of political epistemology needs to come in. Take the assortment of views that go under the general label “climate change scepticism”. First, climate change scepticism might be cited as evidence of the impact of politically-tinged cognitive biases on belief formation — of the way in which our political identities can drive us to reject scientific consensus. Second, it highlights in a particularly vivid way the tension between democracy and technocracy. Maybe it is entirely justifiable for governments to decide on actions to tackle climate change, ranging from the large-scale (e.g. government investment in green energy) to measures that touch on people’s lives (e.g. policies that financially incentivise people to stop using combustion engines or gas boilers), and then go looking for public support for those measures. But it is incumbent on anyone who takes this route to recognise, even if only in private, that it is not exactly democratic to decide on a direction of travel and then try to bring people on board by persuading them of the benefits of that direction of travel.

How should we think about climate change scepticism as a problem in political epistemology? We can start by trying to understand how prevalent it is and what drives it. There are many studies focusing on the US, which tend to find conservative ideology is a particularly strong predictor of climate scepticism there. But there are also studies on China, India, and elsewhere, as well as studies taking a more global perspective, which suggest is that the US case is not straightforwardly generalisable. Conservative political ideology predicts climate scepticism in Western European countries too, but the picture in Central and Eastern Europe is considerably more complicated. Understanding why that is, and what features of those political contexts produce different patterns, is an empirical question with direct implications for the more philosophical questions that political epistemologists want to ask about persuasion, trust, and the rationality (or lack thereof) of “science scepticism”.

Getting the Incentives Right

If that is the right way to think about methodology, it has practical implications for how the field should be structured. Currently, there is far more incentive to work on political issues specific to a very small range of countries, or to work on universal issues while assuming a framing specific to those countries. This shows up in what gets published in prestigious journals, in which examples get taken up and discussed repeatedly, in which research projects attract funding—and the examples with which these projects are framed. Someone working on the epistemological dimensions of political polarisation in the United States is a lot more likely to attract interest in their work than someone working on the same issues in, say, South Korea or Nigeria.

The solution is to change the incentive structure. I don’t have particularly clever proposals about how to do this. But some of what is required is fairly straightforward. We need to change norms around which topics are considered worth discussing, particularly in leading journals. This involves a minimal kind of reflection when reviewing papers, or reading the final published version. Which political context is being assumed, whether implicitly or fairly explicitly? Is it at all reasonable to think that the claims made about, say, polarisation in that context would generalise to other contexts? If they would not generalise, what exactly do we learn from the paper? It also involves actively supporting work that falls outside the usual remit. Finally, it involves—and this is always difficult for academics—a degree of intellectual humility about the extent to which the political context you inhabit has shaped your sense of what the important problems are.

Notice that this is an epistemic argument, not a moral one. There may be reasons of fairness and justice that also speak in favour of broadening the range of political contexts that political epistemologists engage with. But my argument is that it would be an epistemic benefit to the field — that it would produce better political epistemology. The foundational tensions and problems that political epistemologists are interested in manifest differently in different political contexts, and those differences are philosophically informative. A political epistemology that only ever looks at one corner of the world is going to get systematically distorted ideas about the phenomena it studies. Mere parochialism at the level of individuals becomes myopia at the level of the field. The question is whether we are willing to take the steps required to ensure that individual parochialisms add up to something other than collective myopia.

Many think that something has gone wrong with democratic politics.1 Robert Talisse’s Overdoing Democracy is an attempt to explain what exactly has gone wrong. His claim is that we are overdoing democracy. The problem is something he calls political saturation—“the saturation of social life with activities and projects that are overtly organized around the categories and divisions of current politics.” Social spaces—lecture theatres, dinner tables, coffee shops—have become venues in which we perform our political identities. This is a problem because, when we’re not just performing to like-minded partisans, we’re annoying rather than engaging our political opponents, with the result that our political activities are either pointless or actively destructive of the social fabric that holds democracy together. If we want to improve the situation, we need to do something else—something that isn’t politics. As Talisse puts it: “More and better politics cannot be the solution to the problem … because politics is the problem.”

Talisse on Democracy

Talisse’s argument is based on a particular conception of democracy. On what he calls our “ordinary workaday understanding”, democracy is centrally about deliberation—it is not just about voting but what we do in the run-up to voting, like discussing the issues with our neighbours and colleagues, or making the case for our favoured candidate(s) in public forums. This is essentially a down-to-Earth version of deliberative democracy. As Talisse explains:

The central deliberativist thought is that, in a democracy, collective decisions derive their authority from the fact that, prior to voting, each citizen was able to engage in processes whereby he or she could rationally persuade others to adopt his or her favored view by defending it with reasons and offering reasons opposing competing views. According to the deliberativist, then, the democratic ideal has as its core an idea of collective reasoning.

This, as he acknowledges, is a demanding conception of what democracy requires:

the democratic ideal is not realized unless all citizens’ reasons are given a fair hearing in a social process of public deliberation, where individuals are not only willing to question the views of others, but also open to having their own political views challenged (and their minds changed).

The point though is not that we can only view a political system as a democracy if it actually realises this ideal. A democratic system is one in which this ideal is operative. The result is that deliberative democrats are very exercised with finding ways in which to bring democratic practice more in-line with democratic ideals—this is why they talk so much about citizen assemblies, for example, and other venues for more and better deliberation. It is here that Talisse sees a potential problem:

on most deliberative views the entire social sphere is placed within the reach of democratic politics by default. That is, deliberativists tend to see the whole of social space as, at least presumptively, a potential arena for politics, and they regard nearly everything we do as a possible exercise of democratic citizenship.

Deliberative democracy therefore leads to the constant expansion of the political arena. This is especially so if—as will always be the case—there is a gap between democratic politics as it actually is and democratic politics as the deliberative democrat would like it to be. The deliberative democrat’s response is, echoing Dewey, to double down: the only solution for the ills of democratic deliberation is more democratic deliberation.

Deliberative Democracy vs. Democracy

Talisse is very clear on what the case for deliberative democracy is meant to be: it is the best way of dealing with what he sees as a serious challenge to democracy, which is that, while it is premised on the fundamental equality of all citizens, it is still a system where decisions—potentially very consequential ones—that an individual does not agree with and did not support can still be imposed on that individual by the state. The deliberative democrat’s idea is that this sort of coercion is legitimate so long as the citizen on which the decision is imposed had an opportunity to engage in a deliberative process prior to the decision being made.

There is an obvious problem here. Talisse readily accepts that there is a large gap between democratic deliberation as the deliberative democrat envisages it (the ideal) and deliberation as it actually happens (the reality). But if that is so, it is unclear how the ideal is supposed to do any legitimising work. The mere fact that an ideal version of deliberation would legitimise the outcomes of democratic procedures cannot by itself legitimise those actual outcomes, any more than the fact that an ideal version of a trial would produce a legitimate verdict legitimises the actual verdict reached through the haphazard and imperfect procedures of a real trial. The person convicted on the basis of those imperfect procedures is not consoled by the fact that, had the procedures been ideal, the outcome would have been legitimate. And the more imperfect the actual procedures, the less comfort the ideal provides. Given that Talisse regards actual democratic deliberation as inevitably falling very far short of the ideal, it is hard to see what legitimising work the ideal is doing at all.

Setting this aside, there is also the more fundamental problem that deliberative conceptions of democracy fixate on just one aspect of democratic politics. Politics is, in part, a discursive activity; people talk to each other. But politics is also, or rather more so, about power: who has it, who can exercise it, how ordinary people can hold it to account—or at least prevent those with it from dominating them. It isn’t that deliberation is irrelevant: deliberation could be a way of putting checks on power. But it is only a check on power if it connects to actual political decision-making. Absent this, deliberation is, at best, a way in which a community might satisfy itself that, if it had actual political power, it would do more or less what those with power have decided to do. At worst, deliberation, and the venues in which it occurs, become a sort of distraction: there is deliberation, in which ordinary citizens get to play their part, and there is politics, where the decisions are made, and from which ordinary citizens are excluded.

One worry you might have about today’s democracies is precisely that there are far too few connections between the political deliberations among citizens and actual decision-making. Our conversations, our votes, our posts on social media—in fact, almost all of our political activities—don’t make much of a difference to what our political leaders decide. The current war in Iran is a particularly striking example of this: there is simply no way for someone who thinks it is spectacularly ill-advised to translate that into any kind of meaningful action. But there are many other examples you could cite: ordinary citizens lack much in the way of means for holding politicians to account for their economic decisions (tariffs anyone?), healthcare policy, environmental policy, or for that matter any area of policy that matters to people.

If we think about democratic politics as fundamentally about the ability to hold politicians to account, then Talisse’s diagnosis starts to look a lot less compelling. There may be a lot of talking about politics. But that’s not the same thing as genuine political participation.

The difference between talking about politics and politics is akin to that between being a football (soccer) fan, which typically involves a lot of performative talking about football, and actually playing football. There is a lot wrong with football discourse. A lot of it is inane. It can get heated—often quite tribal. Perhaps it would be better if we had a lot less of it. But it would be odd to say that the problem with football discourse is that the fans who engage in it are overdoing football. What they are overdoing is talking about football. Similarly, what political partisans are overdoing is talking about politics, not politics. They might talk less about politics, or talk about it in ways that are less annoying. But that doesn’t change the fundamental point that what they are doing is not, and never was, politics.

The Polarisation Mechanism

Talisse has a nice story about how political saturation and polarisation are mutually reinforcing. The story goes something like this: the modern world produces more opportunities for “political sorting”—we have more choice over where we live and who we associate with, and we naturally gravitate towards people who are like us. This creates a fusion between our social and political identities: politics becomes central to who we are. Once something is viewed in this way, as central to identity, then it takes over everything. It even seeps into our beliefs. The people we surround ourselves with think like us, so they provide constant corroboration for our beliefs, our suspicious, our fears. The result is a peculiar situation: communities are increasingly politically homogeneous (everyone within them thinks more or less the same things) but they develop an intense and mutual hatred for other communities.

I have no doubt that, at least at a high level of analysis, all of this is happening. But I don’t find it plausible that the cause of it is an excess of enthusiasm for democratic deliberation. I also don’t find it particularly plausible that the main driver of political sorting is choice.

Here’s an alternative story: we (not just in the US) live in countries where there are clear “winners” and “losers”. Some people do very well out of modern capitalism. Some do very badly. Many are stuck somewhere in between. Defenders of modern capitalism will argue that, even if there are winners and losers, the losers still—for the most part anyway—do better than they would under any alternative economic system. This may be true—it isn’t the sort of counterfactual that is easy to evaluate, but there aren’t any good examples of better alternatives to hand. Either way, this is not a way of running a society that is conducive to community, solidarity, or any of the traditional means #humans have found to make meaningful connections with each other. If you pit individuals against each other in market competition then, even if most of them end up with more material goods than they would have had otherwise, they end up unhappy, dissatisfied, perhaps even miserable. (It doesn’t make them less miserable to remind them that, if things were different, they would also be miserable but for different reasons).

The winners—and those who aspire to join them—win in part because they become good at navigating this kind of society. They become good at self-marketing, at projecting themselves as the sort of person who might be a winner. They form bonds, albeit tenuous ones, formed for the sake of temporary exigency, with people who they think will help them clamber to the top of the pile. They become good at projecting whatever identities will help them form and maintain these bonds. These bonds are reinforced by the fact that winners tend to live near other winners, in the cities and affluent regions that produce the “innovation” that drives a modern economy. The losers are left to form pseudo-communities around shared grievances. These bonds are also reinforced by the fact that losers tend to live near other losers, in the parts of the country vacated by the winners. Over a few generations, this can become a self-perpetuating cycle: winners have children who become winners, losers children who become losers, because the best way to become a winner is to grow up surrounded by people who are good at doing the things winners do.

The result are communities based around identity projection. In both cases, the identities projected are poisonous, and the communities have only the surface features of genuine communities. They think in similar ways, they do similar things, they have a shared sense of who their friends are and who their enemies might be. But they are far less substantial. They will happily eject someone from the community who deviates from the shared identity, precisely because that identity, rather than any real common stake, is all that holds them together. They are not communities capable of genuine political action. They are founded on shared performance, whether of success or grievance, where these performances are the result of a competitive environment that has progressively dissolved the real communities from which genuine political agency might have grown.

Mouffe on Polarisation and Populism

My alternative story doesn’t really explain why the collapse of genuine communities and the formation of pseudo communities has been accompanied by a rise in political consumerism and the performance of political identity. Perhaps people need to bond over something; but why was that something talking about politics?

Talisse’s answer to this question doesn’t seem particularly compelling. He thinks that the problem is people are applying democratic norms in places that are not appropriate venues for those norms. They are making a misguided attempt to turn the ideals of deliberative democracy into reality. It may be that this is at least partly accurate when it comes to a particular, small, subset of the population: people who are enamoured by the ideals of deliberative democracy, like academics. But I find it implausible that this explains the more general phenomenon. Most people don’t talk about politics because they have, however inchoate, the ideals of deliberative democracy in mind.

A more plausible explanation is that talking about politics, and the performative form that most talking about politics takes, is a substitute for something. But what? Chantal Mouffe’s work on populism and the problems with modern politics is helpful here.

In her work Mouffe agues that the ferocity and tribalism of contemporary political discourse is a predictable consequence of the suppression of genuine political alternatives. The 1990s and 2000s—the era which people who worry the most about modern politics seem to yearn for—was, at least in countries like the US, characterised by a remarkable degree of consensus, at least on what were once the most contentious issues, such as the basic question of what sort of economic system we want. Mouffe’s claim is that, with no genuine left-right conflict on offer, and no real debate about what the future course should be (history had ended), the political energies of both winners and losers lacked a productive outlet. The natural consequence was a transformation of those energies: political activity was no longer about advocating for meaningful changes or involving oneself in the political process but rather about expressing your values, your identity.

For the winners of modern capitalism, politics became just another arena for status competition. This status competition takes on many forms, some of which have been studied by people who are popular with many Substackers. For the losers, it became the only vehicle for expressing anger at the fact they had lost, and the sense of dispossession accompanying that loss. But, without any meaningful avenues for turning this into genuine political action, identity expression became almost an end in itself. Politics has been replaced with talking about politics.

Because they disagree about what has gone wrong with politics, Mouffe’s views about how to fix the problem fundamentally differ from Talisse’s. Talisse thinks the problem is that we are overdoing politics so the solution is to put politics in its place—we should do other things together, things that have nothing to do with politics. Mouffe thinks the problem is that politics has been replaced with talking about politics. We might stop talking about politics so much, or in places where it serves no purpose to talk about politics, but that won’t fix the basic problem, which is that ordinary citizens have no political power, no say in how they are governed, and no way to change a political system in which many of them are condemned to end up losers.

For Mouffe, then, the solution is not a reduction in the intensity of our political discussions but a reorientation of our political energies. We need a real, adversarial politics, a politics that is organised around real questions of power and distribution, with real collective institutions that are capable of reshaping the balance of political power. The question, of course, is how to achieve this, how to get from here to there. One thing that is clear is that we won’t get there just be talking about the need to get there. To think otherwise would, after all, be to confuse politics with talking about politics.

While I am a philosopher, I’m the sort of philosopher who likes to read empirical literature from time to time. One of my interests is science communication, particularly the communication of science that is relevant to political decision-making: climate science, public health, things like that. I got used to reading a lot of articles that look like this:

These articles report on studies—in my non-expert judgement, perfectly good ones—that assess the effectiveness of different strategies for reaching American conservatives who are sceptical about the need to do something about, or indeed the reality of, climate change. These strategies are targeted in the same way that advertising campaigns are targeted; they make certain assumptions about the target audience, and then they craft persuasive messages that are meant to appeal to that target audience. If they “work”, then they lead to a measurable uptick in “pro-climate” attitudes, and therefore might be a valuable tool in the campaign to increase the prevalence of pro-science attitudes. If they don’t, then we need to adopt different tools.

I’ve always thought there was something a bit problematic about this. I used to think the right way to discuss the issue was using the language of “manipulation” and “paternalism”. Advertising is always at least a bit manipulative, especially when it is targeted, so advertising in the name of science shouldn’t be any different. At least it shouldn’t be any different on this score. One difference is that “science advertising” might be justified as being in the service of making our attitudes about science more correct. It is better, you might think, for someone to have a more accurate view about the seriousness of climate change than a less accurate view. Even if it isn’t better for them, it’s better for us, as it makes the task a little bit easier for policymakers who very much want to do something about climate change.

I now think there’s a better way of discussing the issue. Science advertising is basically just propaganda for science. The question then becomes: why is there a popular form of science communication that is engaged in propaganda?

The Strategic Approach to Science Communication

Many advocates of science communication insist that its proper aim is to inform rather than persuade. A particularly clear articulation of this view comes from a short comment piece in Nature, written by a team of leading researchers in risk and science communication associated with the now-defunct Winton Centre for Risk and Evidence Communication at Cambridge. These researchers tell us that the proper aim of science communication is “to design communications that do not lead people to a particular decision, but help them to understand what is known about a topic and to make up their own minds on the basis of that evidence.”

They propose that science communication is primarily about communicating evidence and set out a series of rules for how to do this:

1. Inform, not persuade: The aim of evidence communication is not to change people’s beliefs or behaviours; it is simply to inform.

2. Offer balance, not false balance: The evidence communicator must give a balanced account of the evidence.

3. Disclose uncertainties: The evidence communicator must be honest about what we don’t know.

4. State evidence quality: The evidence communicator must be clear about the quality of the underlying evidence.

5. Inoculate against misinformation: The evidence communicator must anticipate potential misunderstandings and disinformation attacks.

While most of these rules are sensible guidelines, there is a tension. If the rationale for adopting these rules is that doing what they say will enhance the persuasive power of science communication, then it seems more accurate to say that the aim is to persuade by informing rather than to inform rather than persuade.

The fifth rule is particularly problematic. The literature on “inoculation theory” or “prebunking” is quite explicit that the aim is to make audiences less susceptible to misinformation by exposing them to weakened forms of misleading arguments. This goes well beyond anticipating possible misunderstandings. The idea is to shape how audiences will respond to future information by anticipating, in order to block, anything that might “get in the way” of the audience accepting that information. It is hard to see how to justify this if your aim is simply to inform.

When we look at the broader science communication literature, it is clear that the choice of a particular strategy is guided by considerations of its likely effectiveness at “pushing” public attitudes in the desired direction. Take the literature on disclosing uncertainties. There are studies which find that disclosing uncertainties can lead to confusion and lower public support for particular policies and others which report a more mixed picture. The important point is that what is at issue is precisely the effect on persuasive power. But this only makes sense if the aim of science communication is persuasion.

This is clearest in climate science communication, where the central question is how to communicate climate science in a way that secures broader public acceptance of climate change as a real political problem. The two examples from earlier are a perfect illustration of this: the “problem” is one of increasing acceptance among conservatives and Republicans. Researchers recommend techniques including identity affirmation (showing a target group information that affirms their cultural values), narrative framing (using narrative templates congenial to the target group), and pluralistic advocacy (providing advocates that represent a diverse range of values).

I will call this the strategic approach to science communication: an approach that (1) aims to secure broader public acceptance of scientific claims and theories, via (2) the use of evidence-based communication strategies—informed by psychology, communication science, and related fields—to optimise the likelihood of such acceptance.

Dan Kahan, an influential figure in the field, sees the problem of science communication as “understanding the social processes by which the public determines what they see as valid scientific information.” But Kahan doesn’t just want to understand these processes. He wants to understand them in order to “solve” the problem of science scepticism. The strategic science communicator’s task is to identify the communicative strategies that minimise “friction” between scientific claims and public attitudes. This approach is fundamentally strategic because the choice of strategy is guided by whether—by the lights of our best available social scientific evidence—it is liable to achieve the desired attitudinal alignment.

This raises an important question: how should we understand communication practices that systematically select and frame genuine information to achieve predetermined attitudinal outcomes?

How Propaganda Really Works

My account of propaganda is what you might call “sociological”: it views propaganda as a ubiquitous feature of contemporary human societies which has certain functions and serves certain purposes. This contrasts with standard philosophical accounts, which try to provide a definition of propaganda—a set of conditions that cleanly separate propaganda from other forms of communication. The benefit of the philosophical approach is that it provides a certain kind of clarity. But it doesn’t tell us the role that propaganda plays within contemporary societies, and it doesn’t touch on the history of propaganda. Both are important to my argument.

I’ll arrive at my account by considering three figures from the 20th century: Edward Bernays, Jacques Ellul, and Walter Lippmann. Bernays and Lippmann were influential figures in the creation of modern propaganda; Ellul has one of the most developed theories of it. What emerges is an account of propaganda on which it is the “glue” that holds modern mass society together, and—at least in democratic societies—it does this by manufacturing consensus.

In his classic Public Opinion, originally published in 1922, Lippmann frames what he sees as the basic problem: how do we form an opinion about the world outside our own immediate experience of it? His answer is that we rely on stereotypes—interpretive frames that we use to filter and organise the overwhelming mass of information confronting us. But while we can’t do anything but use them, these frames impose order at the cost of distorting our understanding.

You might think Lippmann should therefore defend the value of “ideational heterogeneity”—people having lots of different ideas about the world. But he doesn’t. Instead, he argues that democratic governance requires a high degree of convergence in public opinion. Any democratic society is going to need a tool to manufacture consensus. As he puts it: “That the manufacture of consent is capable of great refinements no one, I think, denies… the opportunities for manipulation open to anyone who understands the process are plain enough.”

Lippmann recognises the need for manufacturing consensus but doesn’t say much about how it’s achieved. Bernays, on the other hand, is very clear that propaganda provides the means—and he’s positively enthusiastic about it. In his 1928 book Propaganda, he writes:

Universal literacy was supposed to educate the common man to control his environment… But instead of a mind, universal literacy has given him rubber stamps, rubber stamps inked with advertising slogans, with editorials, with published scientific data, with the trivialities of the tabloids and the platitudes of history, but quite innocent of original thought.

Bernays was writing at the beginning of the age of mass production. He envisages the mass production of ideas as well as goods. Just as mass production requires consumers ready to purchase standardised products, the mass production of ideas requires a public ready to receive standardised ideas. Bernays viewed the public as passive consumers—ready to receive whichever “rubber stamps” are applied.

What is striking is his understanding of propaganda as a social technology. He defines it as the “consistent, enduring effort to create or shape events to influence the relations of the public to an enterprise, idea or group.” It is a set of techniques for managing public opinion, and like Lippmann, he thinks that mass democracy is only workable if managed by an elite who uses these techniques to engineer consent.

Jacques Ellul, writing in his 1962 book Propaganda: The Formation of Men’s Attitudes, adds two further dimensions. First, he has a larger view of the problem propaganda solves. How do we integrate millions of people into complex systems they cannot hope to understand? How do you coordinate behaviour across a vast population? How do you ensure that members of a society have enough in common for that society to function? His answer is that propaganda provides the “social glue” that holds mass society together. But where Bernays is an enthusiast for propaganda as a means of social control, Ellul is far more ambivalent. He views propaganda as necessary for any functioning modern society, but he is for this reason quite ambivalent about modernity itself.

Second, Ellul argues that democratic political systems require more sophisticated propaganda techniques than authoritarian ones. A democratic state requires popular participation and support. But public opinion is slow to form, unstable, and unpredictable. Propaganda is the means by which democratic governments channel and shape public opinion in ways that support their political ends. Crucially, this often takes the form of rational persuasion:

There is such a thing as rational propaganda, just as there is rational advertising. Advertisements for automobiles or electrical appliances are generally based on technical descriptions or proven performance—rational elements used for advertising purposes. Similarly, there is a propaganda based exclusively on facts, statistics, economic ideas.

The rational propagandist does not simply provide facts, figures and statistics. They provide a sample of relevant facts designed to support the opinions they would like their audience to form. This means there are many ways to “push” the public towards particular patterns of thought and action: presenting facts in ways designed to invoke particular responses; providing biased samples of relevant facts; or only collecting certain kinds of data in the first place.

Taking all this together, I understand propaganda as involving several elements: (1) organised institutional communication (2) aimed at securing behavioural or attitudinal alignment within a population (3) through strategic selection and framing of genuine information (4) where the communicator’s success depends partly on the audience not fully recognising these strategic elements, or at least not experiencing them as manipulative.

This may seem like a slightly unusual way of thinking about propaganda. The standard philosophical account views propaganda as “epistemically flawed” messaging. Sheryl Tuttle Ross’s influential definition is representative: propaganda is an epistemically defective message used with the intention to persuade a socially significant group of people on behalf of a political institution or cause. On this view, propaganda is bad because the propagandist produces messages that are epistemically defective—false, or misleading, or designed to bypass rational deliberation.

My account differs in three ways. First, I’m not trying to capture a commonsensical or pejorative conception. I focus on a way of understanding propaganda that has been historically influential and is central to understanding how modern democracies secure consensus. While the word “propaganda” clearly carries negative connotations, I want to use the concept for analytical purposes rather than simple normative evaluation. If Ellul’s analysis is correct, propaganda is far from an aberration—it is an unavoidable feature of modern society.

Second, I focus on propaganda as targeting behaviours as much as beliefs. If propaganda is the “social glue” that holds society together, it only works if it produces alignment in behaviour as well as belief. It is not enough for the public to believe certain things; those beliefs need to inform their actions.

Third, I understand individual pieces of propaganda within the context of broader campaigns. Ellul distinguishes between “active propaganda” (what we’d ordinarily think of as propaganda) and “pre-propaganda”—the concerted attempt to prepare one’s audience through conditioning, myth-making, or education. The idea is to create a public that will be persuaded by the propaganda that comes later.

Strategic Science Communication as Propaganda

Here is the basic argument. The structure of modern western democracies is fundamentally technocratic: large areas of decision-making have been handed over to experts—central bankers, economists, lawyers, public health experts, scientists. This raises an obvious question: how do we reconcile this technocratic structure with even minimal democratic standards? In what sense is it democratic to hand over decision-making power to experts who we do not elect and are not in any meaningful sense democratically accountable?

One promising answer, developed by the philosopher Elizabeth Anderson, is to view this as a challenge for the institutions that play a central role in our technocratic political system. The basic idea is that legitimacy can be secured via public acceptance and support of the institutions—and the experts who make up these institutions—to whom decision-making power is outsourced.

As it applies to scientific institutions, there needs to be a means by which public acceptance and support is secured. Science communication is one of the central ways in which scientific institutions try to secure this. This applies both in the sense that science communicators seek support for the basic institutions of science—funding bodies, research institutes, universities—and in the sense that they seek acceptance of the basic scientific ideas and theories on which science-informed public policy depends. In climate governance, for instance, the science communicator has the job of securing broad public acceptance of the existence of, and necessity to do something about, anthropogenic climate change.

This project does not necessitate taking a strategic approach. One could imagine carrying it out via deeper engagement with the public of the sort recommended by advocates for engagement models of science communication. I do not want to claim that this sort of engagement doesn’t happen. I simply want to claim that, alongside whatever initiatives there are to bring scientists and the public into dialogue, there is also the strategic approach. And the strategic approach is particularly useful in this context because it allows the science communicator to identify which techniques, in which contexts, and addressed to which audiences, will be effective at securing the desired attitudinal alignment—an increase in pro-science attitudes.

Thus we find central articles in climate science communication with titles like “Improving Climate Change Acceptance Among U.S. Conservatives Through Value-Based Message Targeting” and “Shifting Republican views on climate change through targeted advertising.” While it is possible to combine an approach that views science communication as “targeted advertising” with a desire for greater engagement and dialogue, the obvious issue is that viewing members of the public as objects to be targeted is not conducive to fostering the sort of respect required for genuine dialogue.

What I want to take from this is that the strategic approach to science communication has all four of the key features of propaganda in Ellul’s sense.

First, science communication is always a form of organised institutional communication; the science communicator typically works within an institution, whether scientific or science-adjacent, and is reporting on relevant results of institutional scientific research.

Second, strategic science communication is aimed at securing attitudinal alignment within the public—specifically, wider acceptance of relevant scientific ideas and theories, and more generally, more pro-science attitudes.

Third, the whole point of the strategic approach is to strategically select ways of presenting genuine scientific information that will lead to the desired attitudinal alignment.

Finally, the effectiveness of strategic science communication depends at least partly on being perceived as informative rather than strategic. This is why practitioners often describe their aim as “informing not persuading” even while evaluating techniques based on persuasive effectiveness. The extent to which transparency about strategic aims would undermine effectiveness is an open question, but the field’s self-presentation suggests practitioners believe it would.

My claim, then, is that strategic science communication is a kind of propaganda. Specifically, it is propaganda that plays a crucial role in securing public acceptance of scientific ideas and theories on which the legitimacy of our current technocratic system of governance depends.

This differs from the worry that strategic science communication is manipulative. In a recent paper, the philosopher Mikkel Gerken frames the problem in terms of what he calls “value-based reporting”—science communication that reports a scientific hypothesis in a manner that appeals to the social values of the intended recipients. Some of the strategies I discussed earlier—especially identity-affirmation and pluralistic advocacy—are clear examples. And value-based reporting is clearly manipulative: the audience is treated as an object to be controlled, exposed to inputs designed to exploit aspects of their social identities in order to deliver a desired output.

But framing strategic science communication as problematic because manipulative is to focus on an ethical problem, whereas framing it as propaganda focuses on the political dimension. Specifically, it places science communication within the context of the broader democratic political system in which scientific institutions have their place.

This difference in framing matters because it helps explain why the strategic approach is so attractive to many, even to those who claim their purpose is to inform rather than persuade. Strategic science communication answers to a central problem faced by technocratic political systems: securing public acceptance to maintain democratic legitimacy. For those who want to uphold that system, taking a strategic approach is close to a necessity. Technocratic systems must maintain democratic legitimacy through public acceptance while making decisions based on expertise most citizens cannot independently verify—and strategic science communication provides the most reliable means of securing such acceptance.

It follows that when science communication strategies are manipulative, this is not a failing on the part of individual science communicators. They engage in manipulative practices because doing so advances the aims their institutional role demands. Focusing on whether science communication is manipulative misses something important: the reason why these strategies are often manipulative is because manipulative strategies can be effective propaganda tools.

Science Scepticism

I want to suggest that, in the long term, the strategic approach to science communication risks undermining trust in scientific institutions and so risks leading to an increase in science scepticism.

Here’s the basic argument—or rather a hypothesis I think should be explored: the aim of the strategic approach is to produce consensus. However, in areas of politically contested science, the strategic approach will be unable to produce consensus. As a result, the propaganda employed by the strategic science communicator will not only fail to produce consensus but is likely to lead to reduced trust in scientific institutions and an increase in science scepticism.

A couple of clarifications. First, I am not claiming that the main causes of science scepticism have much to do with the activities of science communicators. It has been thoroughly documented that political ideology is the best predictor of science scepticism in areas of politically contested science. The claim is just that, when an area of science has become politically contested, the strategic approach is liable to make things worse—not that it is the cause of the original problem.

Second, as I will use the term, a “science sceptic” need not reject any particular scientific claim. What matters is that they behave in ways that manifest a lack of trust in relevant scientific claims and institutions—not getting vaccinated, opposing nuclear power or GMO crops, dismissing expert medical advice in favour of alternative treatments. Science scepticism is partly constituted by a lack of trust in scientific institutions. And it can be more or less total: someone might have little trust in public health institutions but more trust in climate science.

The strategic approach views science communication as akin to political communication. The task of the strategic science communicator, like the task of the political communicator, is to identify appropriate communication strategies for different segments of the public. When the public is relatively harmonious—when political divisions are not particularly deep or the cause of serious enmity—it is possible to combine communications targeted to specific segments with “all purpose” communications intended for everyone. This is the situation Ellul’s analysis of democratic propaganda is designed to illuminate: the tools of democratic propaganda are the means by which the communicator can secure general public acceptance and widespread adoption of particular behaviours.

However, when political divisions are deep and the cause of serious enmity, it is no longer possible to produce communications that can reach every segment of society. In such situations, the upshot of propaganda techniques changes. Rather than mobilising the general population in a single direction, they, if they are effective at all, serve to mobilise particular segments. In political communication, the aim becomes to mobilise a large enough group of supporters to win an election or achieve political change. Propaganda becomes a tool of division rather than a way of bringing people together. Far from reducing existing divisions, it exacerbates them because it uses those divisions to mobilise different groups against each other.

In science communication, though, this means the entire project of the strategic approach unravels. Unlike in political communication, there is no benefit to the science communicator in simply mobilising a majority of people in defence of scientific ideas, or in exacerbating existing tensions. The aim is to produce general acceptance of basic scientific ideas and theories—like the existence of anthropogenic climate change—that inform public policy.

There is therefore a basic mismatch between the aim of science communication, which is to produce general acceptance, and what the strategic approach can realistically achieve in contexts where scientific issues have become politically contested. The best that can be achieved is the deepening of pro-science attitudes in groups already inclined towards them. But the cost is likely to be a hardening of anti-science attitudes in groups that are resistant, and a further erosion of trust in scientific institutions among those who perceive themselves as targets of propaganda rather than recipients of information.

I was reading some hardcore analytic epistemology this morning, which is not something I do much of these days. The paper is about the following question: can it ever be rational to both believe that p and suspend judgement on whether it is rational to believe that p? A lot has been written about this and related questions, such as the question of whether it can ever be rational both believe that p and believe that it is not rational to believe that p. (You are supposed to have the intuition that, even if it is never rational to believe that p and that it is not rational to believe that p, it may still be rational to believe that p while suspending judgement on whether it is rational to believe that p).

The standard way of answering this sort of question—in epistemology, but in analytic philosophy more generally—is to make the minimum amount of theoretical assumptions about rationality, about belief, about anything else necessary to frame the question. You do not assume any particular substantive theory of rationality. Or at least you don’t do so explicitly. It may be that you are tacitly relying on a substantive theory, in which case you will be called to task for doing so, either by an irate peer-reviewer or in a response to your paper. You then construct some clever arguments (or arguments disguised as thought experiments1) that lead to the conclusion you want everyone to adopt, or at least consider.

If this works, you get the nice result that it is hard for people to reject the conclusion: you made so few theoretical assumptions that it is hard to deny that the argument carries some weight. This strategy can be remarkably dialectically effective.

I think there’s something odd about this methodology. On the surface level, the oddity is that it requires you to have intuitions about the propriety of saying things like “I believe that p but I suspend judgement on whether it is rational to believe that p”. I’ve never had a good faculty of intuition, but I am so far from having an intuition about this that I can only wonder at those who do.

The deeper oddity is that it is the inverse of what to me seems like the more natural methodology: figure out the correct theory of rationality (and belief, and anything else you need) and then apply that theory to this sort of tricky question. It will tell you what the answer is, and so what your intuition should be. (If you have intuitions that conflict with the theory, well, they are wrong).

Of course, there is a reason why things are this way: the defender of the standard methodology will say that we can only construct a theory of rationality by considering these general questions. They are, if you like, the data around which the theory needs to be constructed.

To my mind, though, this isn’t a particularly compelling response. Even if building a theory of rationality is not like building a physical theory—you can’t go out and do experiments, and no amount of mathematical sophistication will yield testable predictions—it still seems like the standard methodology is wrong. A better way of doing things would be to start by building up a sophisticated theory of rationality, and then apply it to all manner of tricky issues and questions, including the question of whether it can ever be rational to both believe that p and suspend judgement on whether it is rational to believe that p. Different theories of rationality will answer this question, and other questions, in different ways. The hope would be that, at some point, we can crown one of these theories as the “best” theory, and that will tell us what is rational in this and other situations.

I’m someone who is sceptical that this point will ever be reached. Sometimes the data is so meagre that no amount of talk about theoretical virtues can bridge the gap between theory and data. But I still prefer this methodology, where it is all about theory construction, to what seems to be the standard methodology in contemporary analytic philosophy, which is often surprisingly averse to theory. Part of this is, no doubt, an artefact of the peer-review system: you want your paper to get past reviewers, and you have no idea what theory of rationality they favour, so you don’t want to upset them. But it is also an artefact of the requirement to publish relatively small, self-contained units of philosophy to-order. It’s hard to even outline a theory in 8000 words, still less a theory that hasn’t already been outlined somewhere else.

This intersects with another common joke about the propensity of epistemologists to append the word “epistemic” to everything: epistemic autonomy, epistemic paternalism, epistemic self-trust, epistemic sabotage, epistemic exploitation, epistemic appropriation, epistemic objectification, epistemic gaslighting, epistemic manipulation, epistemic silencing, epistemic erasure, epistemic exclusion, epistemic marginalisation, epistemic disenfranchisement, epistemic redlining…. the result can only be described as epistemic exhaustion.

What, exactly, is wrong with this? Why should we be concerned that epistemologists are cataloguing the various types of epistemic injustice, or devising new concepts to analyse? I think the standard answer to these questions is just that it is all a bit banal, perhaps even downright silly. This is often combined with criticism of the ideological homogeneity of contemporary philosophy; epistemic injustice is seen as “woke”. I’m not particularly happy with this answer. I don’t see why cataloguing the different types of epistemic injustice is any sillier than cataloguing the different types of epistemic justification. There may be issues with ideological homogeneity, but they are no more pressing here than in any number of other areas of contemporary analytic philosophy.

My basic take on what is going on here has always been that it is a simple manifestation of the bad incentives driving much of academic research. Young researchers face increasing pressure to publish and the easiest way to publish papers is to jump on board a bandwagon. Your work is liable to get sent to referees who are also riding the wagon, with the result that it is easier to get it published than if you are ploughing away on your own in a more neglected corner of the field. Early in my career I did exactly this, publishing a bunch of papers on epistemic contextualism. This gave me a misleading impression of how easy it is to get your work published, and I didn’t have to suffer the indignity of having my work called out on social media for its banality, partly because that didn’t happen as much in the early 2010s, and partly because people don’t get as annoyed about bank cases as they do about epistemologists talking about social justice.

Central Organising Concepts

I have increasingly come to think there is a more serious problem here though. The problem is with the use of epistemic injustice as what I’ll call a “central organising concept” for thinking about what can go wrong in knowledge production and dissemination. A central organising concept dictates how we think about the core issues and problems in the field within which it does its organising. Fixating on epistemic injustice means focusing more on the reasons why someone might not have their contributions to knowledge production recognised than on what they might have contributed in the first place. It also means paying a lot of attention to what you might call the ethics of knowledge production, particularly to the ethics of interpersonal exchanges. Ethics is important, but what one wants from an epistemological analysis is a focus on the processes by which knowledge is produced, not the niceties of its production.

The rest of this post will be in the service of substantiating what I just said. Let’s start with what I mean by a central organising concept. A central organising concept is a theoretical lens through which a wide range of phenomena are interpreted and understood. It structures inquiry by determining what counts as interesting, what questions are worth asking, what explanations are satisfying. In social science and philosophy, concepts like “power,” “oppression,” and “inequality” often function this way. They’re organising principles that shape entire research programmes, not simple descriptive terms. This is why they can play such a central role even if there is no real agreement on what they mean. Part of the aim of the field is to clarify these central organising concepts.

Epistemic injustice has become a central organising concept in recent philosophical work on knowledge and social (in)justice. Unlike with more nebulous concepts like “power” and “oppression”, there is general agreement that Miranda Fricker’s enormously influential book, Epistemic Injustice, gave us the canonical treatment of what epistemic injustice is. It is a wrong done to someone specifically in their capacity as an epistemic subject—someone who has knowledge they might want to share with others. In Fricker’s treatment, it takes two main forms: testimonial injustice (when prejudice causes a hearer to give a speaker less credibility than they deserve) and hermeneutical injustice (when gaps in collective interpretive resources disadvantage someone in making sense of their own experiences). Subsequent work is more a matter of amending this, or adding further kinds of epistemic injustice (see the images earlier), than questioning Fricker’s basic insight.

The Problem with Epistemic Injustice as a Central Organising Concept

My problem with epistemic injustice as a central organising concept is that, while it is meant to redirect attention towards the contribution made to knowledge production by marginalised groups, what it actually tends to do is one of two things. It either directs attention towards why that contribution wasn’t acknowledged as such, rather than the contribution itself, or it directs attention towards a particular kind of knowledge that is supposedly possessed by members of such groups: experiential knowledge, or knowledge of “what it is like” to be a member of the group in question, or to experience the forms of social exclusion or oppression that members of the group are meant to typically experience.

The problem with directing attention towards why the contribution wasn’t acknowledged is not just that it downplays the contribution; it also serves to obscure the ways in which the groups who have made these contributions react and respond to being ignored. This might involve creating parallel structures that exist outside the institutions that are not listening.

There are two problem with directing attention towards knowledge of “what it is like”. The first is that this is not the only kind of knowledge that marginalised groups produce. Indeed, it may not be the most important kind of knowledge that such groups produce. The aim of early work in feminist science studies was primarily to identify the contributions—typically neglected—of women to science, and the assumptions underlying scientific research that led to those contributions being ignored.

The second and perhaps more fundamental, problem is that it assumes there is one thing that it is like to be a member of any given marginalised group—that there is one thing it is like to be disabled, or black, or working class. The result can be a problematic kind of homogenisation, where the very different experiences people have of, say, being disabled are collapsed into a supposedly universal experience of disability. This flattening can be politically problematic, as when those supposedly universal experiences become the basis of a political programme such as the social model of disability that has a hard time accommodating experiences of disability that don’t fit with the programme—for instance, those who find their chronic disability debilitating by itself, independently of any social exclusion which may result from it.

Patient Activism and Knowledge Production

I want to make what I just said a lot more concrete by drawing on some research I’ve done on patient activism (I wrote about this here). I’m interested in cases where patients—people without formal medical training—have made genuine contributions to medical knowledge. Patient activism is a good place to “test” the value of the epistemic injustice framework because one might expect—and, as we will see, one finds—that there are myriad epistemic injustices that occur in interactions between patients and medical institutions. The examples I discussed in my work include AIDS activists in the 1980s and 90s (as written about in Steven Epstein’s excellent book Impure Science), the Radiotherapy Action Group Exposure (RAGE) in the UK, the French Muscular Dystrophy Association, and the #WeAreNotWaiting movement in the type 1 diabetes community.

I’m interested in these cases for two reasons. First, they do not primarily involve patients sharing their lived experience of illness. AIDS activists weren’t just testifying about what it’s like to have HIV/AIDS. They were using their first-hand experience of “what it is like”, along with an impressive amount of technical knowledge, to critique clinical trial methodologies, point out flaws in research design, and actively participate in the production of medical knowledge. They developed what Steven Epstein calls “lay expertise”—a form of knowledge that allowed them to engage with researchers on technical issues and make genuine contributions to medical understanding. To reduce this to that first-hand experience, to the sort of understanding of what it is to have a medical condition that can only come from having that medical condition, would be to miss what makes Epstein’s case study so interesting and important.

I want to use the other case studies to make similar points. RAGE wasn’t just a group of women sharing their experiences of radiation damage when undergoing cancer treatment. They collected data, identified patterns across hospitals and time periods, and produced analyses that challenged prevailing medical explanations. Like the AIDS activists studied by Epstein, they developed lay expertise—the ability to actively contribute to knowledge and practices within a field.

The second reason I am interested in these cases is how these groups reacted to having their attempts to contribute to medical knowledge and understanding ignored. They erected parallel structures outside of institutional medicine and used those structures to exert political pressure on the institutions to take them seriously. These attempts to exert influence were not always successful; it seems fair to say that, for example, RAGE had less success than Epstein’s AIDS activists. But the point is that these groups took the difficulty in securing an audience for granted—if you like, the impossibility of any kind of epistemic justice was assumed—and behaved accordingly. The #WeAreNotWaiting movement provides an example of this. Members of this community have repurposed and re-engineered continuous glucose monitors and insulin pumps, and have created open-source automated insulin delivery systems. This is a perfect example of a community operating outside traditional research institutions because there is a clear recognition that the traditional institutions just will not do what the community wants them to do. The only alternative is to do it for yourself.

But What Does This Have to Do with Epistemic Injustice?

Whenever I present this research, I’m asked: “How does this relate to epistemic injustice?”I say something like this: epistemic injustice is of course very important. We can imagine—actually we know—that, for example, the members of RAGE were not believed when they reported symptoms from their radiation treatment. This is, of course, a clear example of testimonial injustice.

But this is not the answer I want to give. I want to say that I’m not really interested in how these cases relate to epistemic injustice. I’m more interested in two things. The first is the remarkable fact that people with no formal medical training can make these kinds of contributions to medical knowledge and understanding. This reminds us that there is a difference between possessing important skills and knowledge and possessing the relevant institutional credentials. While it is of course a lot easier to make research advances if you have been provided with the normal training and are working within an institution designed to facilitate research, it is not impossible to do it outside of these contexts. This often gets lost in the handwringing about the crisis of trust in expert institutions. There are problems with expert institutions and reacting to those problems by creating something outside of those institutions is not in itself to reject expertise or expert knowledge. To assume that the credentialed experts have a monopoly on knowledge is simple credentialism.

The second is that these groups did not react to being sidelined or ignored by traditional research institutions by trying to identify a kind of knowledge—experiential knowledge about what it’s like to be a patient—that they could credibly claim that only they possessed. Instead, they established research collectives that functioned analogously to traditional research communities. They created structures for training members, developed norms governing knowledge production and sharing, and established venues for knowledge dissemination. They made substantive contributions to medical knowledge and, in some cases, developed new medical technologies.

What the epistemic injustice framework wants to do is direct our attention to the wrongs done to these groups, and to highlight that members of these groups possess a kind of expertise—first-personal experiential expertise, knowledge of what it is like to be a member of one of these groups—about which their credibility cannot be questioned. This may matter for certain purposes, but it risks obscuring what these groups actually accomplished, or at least were trying to accomplish. It frames their knowledge production primarily as the identification and attempted correction of injustice rather than as a positive achievement in its own right.

What We Should Focus On Instead

If we shouldn’t focus on epistemic injustice, what should our central organising concept be instead ? I’m not sure we need a central organising concept. But here are some ideas about what we could focus on, at least as social epistemologists trying to understand the ways in which knowledge is produced.

First, we should focus on knowledge production itself—on the actual processes by which groups operating outside of the normal expert institutions create new knowledge, develop expertise, and contribute to collective understanding. This means paying attention to the social structures that enable this knowledge production: the research collectives, their training structures, the norms and practices that govern inquiry within those collectives.

Second, we should focus on the conditions under which alternative forms of expertise emerge and gain traction. When do patient research collectives succeed in producing valuable knowledge? What forms of cognitive scaffolding do they need? How do they interact with traditional research institutions? These questions are more productive than simply cataloguing instances of epistemic injustice.

Third, we should be more attentive to the diversity of knowledge types. Experiential knowledge matters, but it’s not the only kind of knowledge that matters. Technical knowledge, methodological knowledge, practical knowledge—all of these can be produced by communities operating outside traditional institutions.

Finally, if—like me—you find the epistemic injustice programme increasingly looks like a research programme that has run its course and has started to degenerate, we should be looking for the dimensions of knowledge production that are not amenable to analysis using the framework of epistemic injustice. These tend to be material and structural rather than interpersonal and ethical: questions about access to resources, control over research infrastructure, and the organisational forms through which knowledge gets made. Epistemic injustice directs our attention to the ethics of testimony and interpretation. What we need instead is something closer to a political economy of knowledge—an analysis of who has the capacity to produce knowledge, and what is required to have that capacity, not just who gets believed.

At the start of the year I told myself that I would read more non-fiction, particularly the sort of non-fiction that is meant to appeal to someone with diverse academic interests (you know, the sorts of things someone like Dan Williams might recommend you read). Safe to say, it didn’t go terribly well, but I managed a few:

1. Hannah Arendt, On Violence

2. Edmund Burke, Reflections on the Revolution in France

3. Alexander Douglas, Against Identity

4. Frantz Fanon, The Wretched of the Earth

5. Michael Oakeshott, The Politics of Faith and the Politics of Scepticism

6. Richard Rorty, Pragmatism as Anti-Authoritarianism

7. Cory Wimberly, How Propaganda Became Public Relations

The Arendt, Burke, Fanon and Oakeshott were all for a class I was teaching on philosophy and social change. (I should add that of course I read everything on the syllabus—but these were complete books I had not read before). I can’t think of a class I have more enjoyed teaching. The idea behind the class was to look at different theories/understandings of social change—roughly, conservative, radical and liberal—and to relate those theories to concrete historical events (the French revolution, the Haitian revolution, the rise and fall of communism, colonialism and decolonisation).

I’d never read Burke before; I came away impressed by his rhetorical skill but a little disappointed in him as a thinker. Perhaps the best thing about having read Burke was that I could properly appreciate Mary Wollstonecraft’s response, the pamphlet “A Vindication of the Rights of Men”. Here’s a flavour (most of it is like this):

Man preys on man; and you mourn for the idle tapestry that decorated a gothic pile, and the dronish bell that summoned the fat priest to prayer. You mourn for the empty pageant of a name, when slavery flaps her wing, and the sick heart retires to die in the lonely wilds, far from the abodes of men. Did the pangs you felt for insulted nobility, the anguish that rent your heart when the gorgeous robes were torn off the idol human weakness had set up, deserve to be compared with the long-drawn sigh of melancholy reflection, when misery and vice are thus seen to haunt our steps, and swim on the top of every cheering prospect? Why is our fancy to be appalled by terrific perspectives of a hell beyond the grave? – Hell stalks abroad; - the lash resounds on the slave’s naked sides; and the sick wretch, who can no longer earn the sour bread of unremitting labour, steals to a ditch to bid the world a long good night – or, neglected in some ostentatious hospital, breathes his last amidst the laugh of mercenary attendants. Such misery demands more than tears – I pause to recollect myself; and smother the contempt I feel rising for your rhetorical flourishes and infantine sensibility.

I was far more at home with Oakeshott. I plan to write more about this in the months to come, but the abiding impression I got from reading Oakeshott was that he was making a real attempt to take a step back from the political scene in order to try and understand broader dynamics and processes. He does this in a conservative way, of course—complete detachment and neutrality is impossible—but so few writers on politics, even supposedly serious ones, even make the effort.

I had read some Fanon before, but never the whole of Wretched of the Earth. I must admit some of the later chapters, which often verge on programmatic communism for the newly liberated postcolonial world, bored me a bit. It is of course the first chapter or so, on violence, that attracts most readers—and repels many. I must admit to not really seeing what is so controversial in it. The observation that extremely violent treatment is liable to evoke an extremely violent response, and that the extremely violent response may be experienced, at a psychological level, as cleansing, and at a higher level as some sort of re-assertion of one’s dignity (which for Fanon was never lost in the first place, only dormant), strikes me as clearly accurate.

Arendt (who I generally have a slightly troubled relationship with) seems to me, in On Violence, to understand Fanon perfectly well—as she is at pains to say, better than Sartre seems to have, judging on his preface to Wretched of the Earth. She agrees with Fanon that, if you try to control people with violence, they are liable to strike back with an equal amount of violence as soon as they get the opportunity to do so. The point she wants to make about violence is astute and remarkably clever: it cannot ever be the basis of genuine political power because genuine political power requires the capacity to move people to do things without coercion. (There are some subtleties here—what about subtle forms of coercion?—but clearly there is nothing subtle about coercion via violence). This, among other things, allows her to explain why revolutions can happen so suddenly (the old regime had a remarkable capacity for violence but not much genuine power, so the minute anyone tried to topple it, it fell).

As for the rest, Douglas’s book is excellent; I can’t recommend it highly enough—or commend him highly enough for absolutely resisting the temptation to say anything about current politics in it. I enjoyed the Rorty, but the basic thesis—that pragmatists should reject the notion that our thoughts should try to be true to an external reality because that would represent a kind of authoritarian control over those thoughts—strikes me as, frankly, a bit mad, though perhaps a useful way of understanding what really drives someone like Rorty. I wrote about Wimberly’s book here. All I will say about it right now is that it is a lot more interesting than a recent book written about propaganda by another, more famous, philosopher.

So onto the fiction:

1. Jane Austen, Emma

2. Jane Austen, Sense and Sensibility

3. James Baldwin, Go Tell it on the Mountain

4. Anne Brontë, The Tenant of Wildfell Hall

5. John Buchan, The Thirty Nine Steps

6. Anton Chekov, Tales of Russian Life

7. Liu Cixin, The Three-Body Problem

8. Fyodor Dostoyevsky, Devils

9. Mavis Gallant, Green Water, Green Sky

10. James Kelman, A Chancer

11. Andrey Kurkov, Penguin Lost

12. William McIlvanney, Strange Loyalties

13. William McIlvanney, The Big Man

14. Viktor Pelevin, Generation P (aka Babylon, Homo Zapiens)

15. Leonardo Sciascia, To Each His Own

16. WM Sebald, Austerlitz

17. Nan Shepherd, The Living Mountain (ok not technically fiction but it fits better on this list than the previous one)

18. Vladimir Sorokin, Day of the Oprichnik

19. Ivan Turgenev, Fathers and Sons

20. Evelyn Waugh, Decline and Fall

I enjoyed most of these. The only one I didn’t like at all was the Gallant, which was a little too transparently “writerly” for me; a lot of craft, but a lack of genuine emotion. Go Tell it on the Mountain displays the same sort of craft, but combined with real passion. The Tenant of Wildfell Hall has some unfortunate structural flaws but also a clarity of purpose that keeps the whole thing together. Penguin Lost is markedly inferior to Kurkov’s earlier Death and the Penguin (it is a sequel) but it’s diverting enough. I don’t want to judge The Three-Body Problem too harshly (and it’s a great premise) but some of the writing, at least in translation, is bad even for science fiction writing.

A few of these names might be unfamiliar: Kelman, McIlvanney and Shepherd are all Scottish writers. Kelman and McIlvanney are relatively contemporary (in fact Kelman is still alive and writing); Shepherd died in the early 80s. Kelman is a writer in the 20th century European existentialist tradition, but his characters speak in a Glaswegian dialect. The closest comparison I can think of would be someone like Thomas Bernhard (but, as I said, more Glaswegian, and with more swearing). McIlvanney is, as best I can judge, a bit less critically acclaimed than Kelman but a more entertaining read. The Big Man is about one man’s attempt to reclaim his own self-worth, and the self-worth of the small mining town he comes from, by taking part in a bare knuckle fight organised by some drug dealers. If I were to blurb it for the NYT I would say it is an exploration of toxic masculinity, but that would make it sound like it is something frivolous, or worse “woke”, which it isn’t. Shepherd only published The Living Mountain, which is one of the best examples of British nature writing, and a series of books about the lives and travails (mostly travails) of rural communities, particularly women in those rural communities, in the early 20th century. I wrote about one of these books, The Quarry Wood, last time.

There’s quite a bit of Russian literature in there. This wasn’t deliberate. Everyone knows Chekov, Dostoyevsky and Turgenev. Pelevin is, I gather, very famous indeed in Russia, but perhaps less so outside of it. Generation P is not a complete artistic triumph, but part of the reason for this is that it is trying to describe a future with which we are now intimately familiar but using a vocabulary that is not quite adequate to the task (William Gibson’s Neuromancer has a similar problem). The basic premise is that a cabal of businessmen, gangsters (also known as businessmen) and TV executives have conspired to create a bunch of fake politicians who have staged political battles that serve as cover for the real battles going on within the cabal. Day of the Oprichnik is regularly listed in articles on recent Russian classics. I must admit I struggled a bit with it before reading the Wikipedia page, which explains, in great detail, the Russian novel that it is largely a parody of, and that provides the backstory that is only hinted at in the book.

What else? Decline and Fall is hilarious; a great way to spend a wet and dull afternoon. To Each His Own isn’t funny, exactly, but it has a kind of dark comedy (it’s the story of a murder, but the identity of the killer is clear, and we’re in a fatalistic world where murders like this just happen sometimes). I like Austen but I can’t get down with the people who really love Austen. Everything is too neat, too tidy. All the characters’ foibles (and there are a lot of them) are largely harmless. You get the sense that nothing too awful can happen in an Austen novel. I prefer the quiet madness of Fathers and Sons, or the sheer lunacy of Devils. Lastly, Austerlitz was the best thing I read this year. Everyone should read it.

One thing I don’t like, though, is selective cynicism. What do I mean by selective cynicism? An example will be helpful. Because I don’t want to pick on anyone in particular—and I think selective cynicism is extremely common, so picking on one person would be unhelpful—I’ll make it fictitious. Let’s imagine a character: Professor X. Professor X is a proficient social media user (from what follows you will be able to infer that Professor X most likely uses X). They constantly post things like this (their wording is more outraged and arch than my fictitious renderings):

• Academics fashion their research around fads and trends.

• Academics don’t “speak up” due to the fear of being cancelled.

• The main reason people, especially academics, sign petitions is because of the fear of being censured if they don’t. (And the hope of being praised).

• Academics praise work that advances political conclusions they agree with.

• Academics want to hire “their people”—other academics who think like them, especially when it comes to politics—and they have erected an enormous bureaucracy to enable this.

• Academics don’t really believe the progressive orthodoxies they spout; they are “luxury beliefs”.

You get the idea. Of course, by “academics” they don’t mean all academics. They exempt the brave truth-tellers, the academics who avoid the fads, who speak their minds, who are willing to engage with anyone.

As a cynic, I don’t really disagree with any of these claims. (Though I doubt the utility of the category “luxury belief”. It’s either a simple partisan weapon or it can be used against everyone likely to use the category to diagnose failings in others). I also don’t want to disparage the idea that there really are people out there who are willing to go against the grain, stick their necks out, and give voice to uncomfortable truths. (I can’t resist adding though that there are far fewer people who fit this description than there are people who present themselves as fitting it. As always, it’s easier, and about as financially rewarding, to appear to be something than to actually be it).

My problem is that the Professor Xs of this world apply this cynical way of thinking of human (well, academic) behaviour selectively. They apply it to their political enemies but not to their friends and allies. The heterodox academics who present themselves as truth-tellers, courageously speaking out against stifling orthodoxy? They're motivated by pure intellectual honesty, naturally. The people who share their contrarian takes widely on social media? They're the rare souls willing to follow evidence wherever it leads. The academics who hire these contrarians or praise their work? That’s just merit recognising merit. The Professor Xs will dissect the base motivations behind research they disagree with, but when their ideological allies publish work that supports their preferred conclusions, suddenly it's rigorous scholarship rather than confirmation bias. They'll mock academics for chasing fashionable topics while remaining curiously silent about how "anti-woke" scholarship has become its own cottage industry. There’s money to be made for the member of the elite who is willing to “tell the truth” about the other elites.

If you're going to be cynical about human motivations—and I very much think you should be—then you have to be cynical about all human motivations, including your own. Otherwise you're not really a cynic at all. You've just found an intellectually sophisticated way of rationalising your tribal allegiances. Which, when I think about it, is exactly what I would expect Professor X to do. If you think we can explain much of human behaviour in terms of status competition, then it is no surprise that we find a section of the elite—the “counter-elite”—trying to gain status by cataloguing all the failings of the elite class. Of course, I need to follow my own advice and apply this to myself: what is this but an attempt to gain status by pointing to the cynical applications of cynicism by the new counter-elite? Unfortunately for me, at this point the status competition has become so incestuous that there’s probably not much money to be made in the attempt. If I’m lucky it will gain me a few new followers though.

Subscribe now

I've spent a lot of time recently thinking about propaganda, democracy, and many of the problems of our age. This interest was partly sparked by reading Jacques Ellul's Propaganda: The Formation of Men's Attitudes, which offers a compelling account of propaganda as a necessary feature of technological society. But it was also motivated by the sense that a lot of contemporary discussion of propaganda misses something crucial about how it actually works.

Cory Wimberly's How Propaganda Became Public Relations attempts to fill this gap by offering what he calls a "genealogy" of propaganda. His central thesis is that:

Propaganda is not different from public relations except in its name; public relations and propaganda name the same activities, the same rationalization of those activities, and even the same personages. They are the same except that public relations is a kind of doubling of propaganda, in that the term 'public relations' is propaganda for propaganda.

As a claim about the common understanding of the word “propaganda”, this thesis is not particularly attractive. When we think of propaganda, we think of Goebbels, socialist realist posters, and the like. We don’t think of corporate PR. But Wimberly is offering a genealogy, not a conceptual analysis. He argues that, by tracing the historical development of public relations in early 20th-century America, we can understand propaganda not as epistemological manipulation (false beliefs, misinformation, and so on) but as an apparatus for constituting subjects—for making people into the kinds of people who will act in ways that serve corporate interests.

The book is ambitious, combining archival research into early public relations with a Foucauldian analysis of power and subjectification. It also has a lot to say about democracy, and the extent to which we (well, Americans) live in one. Wimberly's argument is that propaganda represents a fundamental challenge to liberal democracy. But the reason it represents a challenge is not because it deceives people or leads them to act against their self-interest. It is because it transforms us. As Wimberly puts it:

This work, instead of looking at propaganda as papering over reality with false belief, looks at propaganda as having an important role in creating it.

The propagandist's aim is not deception but creation. The propagandist is engaged in the manufacturing of subjects. The power to manufacture subjects is enormous, and if corporations (Wimberly seems to treat government as an extension of corporations) have that power then there is no meaningful sense in which we live in a democracy, whatever ideals our corporate elites may profess to hold.

The Genealogy of an Apparatus

Wimberly's genealogical method details the emergence of what he calls “apparatuses of propaganda”—the component parts of a propaganda, the relationships between them, and the urgent needs that they met.

The genealogy begins with World War I and the success of wartime propaganda in "remaking" the American people into enthusiastic supporters of the war. Edward Bernays, who would become one of the most prominent theorists and practitioners of public relations, observed this success and drew an obvious conclusion:

I decided that if you could use propaganda for war, you could certainly use it for peace. And 'propaganda' got to be a bad word because of the Germans using it, so what I did was to try and find some other words, so we found the words 'public relations'.

This was more than just a rebranding exercise. According to Wimberly, early 20th-century corporations faced a crisis of legitimacy. Strikes, regulation, boycotts, and muckraking journalism had created what he describes as "problematic relations between corporations and their publics”. The traditional liberal response—defending corporate rights to non-interference—was no longer sufficient. Instead, corporations recognized that they needed to create a more cooperative—that is, more compliant—public.

This governance took a particular form, drawing heavily on the emerging field of crowd psychology. Thinkers like Gustave Le Bon had argued that people in groups don't act as rational individuals but as irrational crowds, susceptible to manipulation by skilled leaders. This provided a theoretical foundation for a new approach to corporate-public relations. Rather than simply defending their actions, corporations would reshape the public to want what corporations wanted them to want.

Wimberly's account of this transformation is detailed and grounded in the archive materials of prominent propagandists, like Bernays and Ivy Lee. He shows how propagandists like Bernays and Lee developed sophisticated techniques for "engineering consent”. Take, for example, Bernays' famous campaign to get women to smoke Lucky Strikes. Rather than simply advertising cigarettes, he positioned smoking as an act of women's liberation, recruiting women to break laws against women smoking in public during the 1929 Easter Day Parade in New York City. The key insight Wimberly draws from such campaigns is that they weren't just about selling products. Bernays wanted to sell cigarettes by constituting a new subject: the liberated woman smoker. As Wimberly puts it, "the aim was to produce the relationships that corporations wanted" which required significant changes to the "subjectivity of the publics so that their values, ambitions, desires, and sensibilities were allied with corporate prerogatives." That might have meant smoking cigarettes, buying a new car every few years, or voting for politicians who would protect the power and influence of corporations.

Propaganda and Liberalism

One of Wimberly's most important contributions is his analysis of how propaganda emerged as a response—and challenge to—liberal democracy. The propagandists weren't simply trying to work within liberal institutions; they were developing techniques to circumvent them.

The basic liberal ideal, as Wimberly understands it, assumes a society of autonomous individuals who are capable of knowing their own interests and making rational decisions about them. For the liberal, this is an ideal; it is acknowledged by everyone that people are sometimes wrong about their own interests, and they don’t always act rationally in the pursuit of them. But propagandists, drawing on crowd psychology, rejected the assumption entirely. This, combined with the complexity of modern society, led propagandists to a radical conclusion:

Rather than view the individual as the one who best knew its interests and the state as unable to know those interests, now it was the publics who were unable to understand the modern context and see what was in their best interests and the corporations who [were] best able to direct them on the path of progress and civilization.

Thus, for Wimberly propaganda, and public relations, represented a way for elites to maintain control without directly challenging democratic institutions. As he explains:

Propaganda thus proposed governing the desires of the publics so that they might appear to respect the rights of the liberal subject ... but nonetheless subjecting the publics to domination that lied outside of the liberal framework's [capacity] to grasp and understand.

The propagandists were explicit about this, at least in their writings. Bernays is quoted frequently throughout the book, and for good reason; he wasn’t shy about expressing his anti-democratic elitism. In one striking passage Bernays tells us that, while "it might be better to have ... committees of wise men who would choose our rulers, dictate our conduct," we don't have such a system. Instead, "ours must be a leadership of democracy administered by the intelligent minority who know how to regiment and guide the masses." This is, in essence, a defence of managed democracy; a democracy that is managed in such a way that the people (“the masses”) make the choices that the elites want them to make. The task of the propagandist, as Bernays saw it, is to do the managing.

The Constitution of Subjects

Another of Wimberly’s important contributions is his Foucauldian analysis of how propaganda works through the constitution of subjects rather than simple deception and misdirection. Drawing on Foucault's concept of government as “the conduct of conduct," he tells us that:

Propaganda is a response to free conduct and an attempt, not to dominate it via violence or physical constraint, but to conduct it in its freedom towards the desired outcomes.

The propagandist wants people to behave in a certain way, but they want them to choose to behave in that way, to—as it appears to them anyway—freely do what the propagandist wants them to do. The basic aim of propaganda is to constitute groups of people (“publics”) who desire to act in the ways the propagandist wants—buy cigarettes, vote for pro-business politicians, contribute their time and money to the war effort. The thought is that, if you can create a public who wants to do what you want them to do, then coercion and other forms of domination aren’t required.

But how is this meant to work, exactly? Why view this as the constitution of a new subject, as opposed to the attempt to influence an already constituted subject to act in particular ways, ways that the propagandist wants? In the background Wimberly is working with a picture according to which what we are is simply defined by what we do. He explains:

Once a subject is defined by its conduct and not by appeal to a subjective essence or nature, then changes in conduct are changes in the subject by definition. Like a painting, in which a change in a few brush strokes on its surface would produce a new work of art because the surface and the depth, the essence and its expression, are collapsed, so too with the subject are alterations in its conduct productive of a new subjectivity. The hard metaphysical distinction between subjective surface and depth is dissolved and instead become a matter of degree rather than difference in kind.

Wimberly is imagining this sort of objection to his proposal: what the propagandist does is simply take an already constituted human subject and push them to act in ways that the propagandist desires (buy their product, vote for their chosen candidate). This doesn’t re-constitute the subject in any meaningful sense. It simply pushes their existing wants and desires in a particular direction. The skilful propagandist harnesses our innate desire to lord it over others, or—if you are of a less pessimistic disposition—our innate tendency towards benevolence and compassion to help in a way that the propagandist desires.

But Wimberly is sceptical that we have any innate desires or tendencies. We don’t come into the world pre-formed. We are formed by the world, and in particular we are formed by the apparatuses of propaganda that we encounter. From this perspective, propaganda doesn't simply manipulate pre-existing subjects; it creates new kinds of subjects—consumers, employees, voters—whose desires and behaviors align with corporate interests.

It is perhaps easiest to see what Wimberly is getting at by considering one of his examples: the construction of the consumer. In the early days of the 20th century, Wimberly tells us, corporations were faced with a problem: how to find a market for the goods they were producing in increasingly large quantities? Bernays was clear about the challenge:

While, under the handicraft or small-unit system of production that was typical a century ago, demand created the supply, today supply must actively seek to create its corresponding demand.

People weren’t used to buying the things they needed; they were used to making them, or borrowing them from someone else who had made them. This was a problem for someone who wanted to sell a product. Bernays’ solution, and the solution of public relations firms across America, was not just to advertise products but to create people who would want those products—to manufacture the consumer as a type of person. This, the thought went, would solve the problem of insufficient demand for the products that the growing industries were producing.

Wimberly is talking about the formation of what we would now call “consumer culture”. It is tempting for many—perhaps particularly the sort of person who reads a Substack like this one—to think they are exempt from the pressures of consumer culture. But Wimberly emphasizes that these processes target masses, not individuals:

An individual might feel as an individual that she is critical and little impacted by propaganda: after all, does not she know that it is all smoke and mirrors? However, what we are judging is not what the individual thinks but what the public does.

The propagandist isn’t interested in what anyone thinks about the choices they make as a consumer. They are interested in what they do, and in particular what they buy:

An individual may think whatever she wants, but if in her collective actions ... she acts in the collective fashion the propagandists create (e.g., watches the television, buys a nice midsized SUV, or completes the university administration-mandated course assessment work), then what do the rebellious thoughts of the individual matter to the propagandist who is measuring TV viewership, automobile sales, or how well a university serves its core customer base?

As Wimberly pithily remarks:

If a radical still buys the dress, movie ticket, or 12-piece dinner set, what does it matter that it is a radical who buys it when there is no room for that radicalism to manifest itself in the public?

What matters to the propagandist is that you behave in the ways that they want, not that you understand yourself as the unwitting instrument of the propagandist. Indeed, propaganda wouldn’t be effective at all if you had to view yourself as its instrument in order to be influenced by it.

Methodological Limitations

Wimberly offers some valuable insights into the historical development of propaganda, and into the tensions between propaganda and democracy. There are however some significant problems with his approach which, I think, limit its explanatory power and call into question some of his core claims about propaganda.

The first problem is methodological. Wimberly's genealogy relies heavily on the archives of prominent propagandists—people like Edward Bernays and Ivy Lee. This approach has the virtue of taking seriously how practitioners understood their own work. But it faces two significant problems.

The first is that propagandists may not have been entirely honest in their writings. As Wimberly himself notes, public relations is "propaganda for propaganda"—part of its function is to legitimize and obscure its own operations. When Bernays writes about the power of propaganda—its ability to reshape subjectivities, to create whole new categories of human beings—we should at least consider the possibility that this is itself a form of propaganda designed to enhance his professional standing and that of his field. Above all else, a successful propagandist needs to be able to sell themselves to their audience. This means that the claims of the propagandist about their successes should be taken with a handful of salt.

Second, even if we accept that the claims of propagandists about what they were doing weren’t entirely disengenious (they bought their own bullshit), it’s a further question whether they had any real insight into why what they were doing worked, to the extent that it worked at all. Take the creation of consumer culture. The way Wimberly tells it, this was largely the result of the decision of propagandists to create a new form of subjectivity. But the creation of consumer culture surely resulted from a complex interaction of technological, economic, and social changes that go far beyond the specific techniques developed by public relations professionals. The propagandists may have been riding broader historical trends rather than directing them.

These methodological concerns are compounded by the geographical scope of Wimberly's analysis. His focus is largely on propaganda in the United States, with some material on its European intellectual roots in crowd psychology. But this means his account is necessarily incomplete as a genealogy of propaganda elsewhere in the world. Different societies at different times have developed their own relationships between power, knowledge, and public opinion. Propaganda is hardly a uniquely American (or European) phenomenon. A genealogy that focuses primarily on American corporate public relations cannot capture the full range of ways that modern societies have organized these relationships. What about propaganda in Europe, South America, Africa or Asia? Did it develop in such close proximity to corporate public relations? If it didn’t, what becomes of Wimberly’s attempt to conceptually connect propaganda with public relations? There is a difference between a genealogy of propaganda, which is what Wimberly claims to provide, and a genealogy of a particular species of propaganda—the propaganda of corporate America in the 20th century.

The Limits of Subjectification

The second problem relates to Wimberly's claims about the extent to which propaganda constitutes subjects. While his Foucauldian framework provides useful tools for analyzing how people's conduct is shaped, he may overestimate both the malleability of human nature and the power of external influences to reshape it.

Wimberly's account requires a strong version of social constructionism about human subjectivity. Human nature presents no constraint to the propagandist; it is treated as entirely elastic—as a lump of clay that can take on whatever form the propagandist might wish. This is, to say the least, questionable. You might think, for instance, that people have a considerable capacity for resistance, adaptation, and independent judgment even under conditions of extensive social pressure. You might think that humans have certain innate tendencies, implanted by evolution. You might also think that people differ in these capacities and tendencies: some people are more resistant to social pressure than others, some people are more benevolent than others, and so on. An analysis like Wimberly’s completely ignores these and other possibilities.

Anyone who is familiar with criticisms of Foucaldian projects like Wimberly’s will be familiar with these sorts of worries. They may also have already anticipated a deeper worry, which is that there is a problematic sort of circularity in Wimberly’s argument. He claims that we don't recognize the effectiveness of propaganda because it has made us the kinds of people who passively accept its ubiquity. But this makes his theory somewhat unfalsifiable—any evidence that people have retained a form of subjectivity outside of the influence of propaganda can be dismissed as a failure to understand how thoroughly they've been shaped by propaganda.

The Necessity of Propaganda

it is instructive to compare Wimberly’s view of propaganda with another view that I have written about extensively: Jacques Ellul’s. While Wimberly does discuss Ellul at a few points in his book, he doesn’t address what to me seems to be the most interesting difference between them. Ellul views propaganda as a necessary feature of modern technological society rather than a tool wielded by particular interests. It is necessary because, absent propaganda, there is simply nothing else that might hold modern society together. Wimberly may also be said to view propaganda as a necessary feature—indeed, as the bedrock—of modern corporate America, but he seems to think that it is also a deeply contingent feature. Things could have been very different. In contrast, for Ellul, there is a sense in which things could not have been different; the modern world is simply unimagineable without propaganda, so the only way in which we could do without it would be by not living in the modern world.

I suspect Ellul is right about this. This point needs to be handled with care, though. From Wimberly’s perspective, Ellul is a sort of apologist for propaganda. He seems to understand Ellul as saying something like this: Modern societies are simply too complex for individuals to understand through direct experience. We need interpretive frameworks to make sense of the world, and these frameworks must be produced and disseminated by institutions with the resources to do so. These are the institutions that produce propaganda, and the modern world can be understood as a sort of propaganda “arms race", where rival governments try to keep control at home and gain influence abroad by engaging in it. You might find some forms of propaganda distasteful, but it’s not something that modern society could do without.

I don’t think this is the right way to understand Ellul’s view. The basic problem is that it misinterprets Ellul’s attempt to understand why propaganda is an essential feature of modern society as an attempted vindication of modern society and the role that propaganda plays in it. This misses the degree to which Ellul is deeply ambivalent about modern society precisely because of the role that propaganda plays in holding it together, which he thinks is simply incompatible with democracy. Here are some passages from Ellul that I think capture the ambivalence:

Once democracy becomes the object of propaganda, it also becomes as totalitarian, authoritarian, and exclusive as dictatorship.

The democratic State, even if it does not want to, becomes a propagandist State because of the need to dispense information... It is, in effect, a State that must proclaim an official, general, and explicit truth. The State can no longer be objective or liberal... It can no longer tolerate competition, because a State that assumes this function no longer has the right to err; if it did, it would become the laughing stock of the citizenry, and its information would lose its effect, together with its propaganda. For the information it dispenses is believed only to the extent that its propaganda is believed.

Democracy is not just a certain form of political organization or simply an ideology—it is, first of all, a certain view of life and a form of behavior... But if democracy is a way of life, composed of tolerance, respect, degree, choice, diversity, and so on, all propaganda that acts on behavior and feelings and transforms them in depth turns man into someone who can no longer support democracy because he no longer follows democratic behavior.

These passages don’t sound like an apology for propaganda. They read instead like an indictment of it. Ellul, in short, thinks that propaganda is incompatible with democracy and freedom. But, because he thinks that propaganda is unavoidable in modern society, he seems forced to conclude that modern society is also incompatible with democracy and freedom. Hence the ambivalence: he essentially has an argument against modernity, which breeds a deep form of pessimism.

The problem is not just that Wimberly seems to misunderstand what Ellul was trying to do. The problem is that, if Ellul is right, resisting propaganda is going to be even more difficult than Wimberly already thinks it will be. Wimberly finds a sort of hope in the fact that, on his analysis, propaganda might not have existed in its present form. He writes that:

The aim of this text is not to reclaim the truth of the human essence so that an ideal political regime can be constructed or recovered. Its aim is to write a genealogy of propaganda so that we may better see just how contingent and fragile our present is.

Wimberly acknowledges that simply recognizing this contingency and fragility is not enough. Propaganda needs to be actively resisted, which means fighting back against the subjectivities it has created—against being the people that propaganda needs us to be if it is going to function. This leads to a defence of a radical form of democracy:

Democratizing the subjectification and government of propaganda means involving people in their own self-formation as individuals and groups as much as possible—and not just when they come up with the 'right' outcomes.

The sentiment is laudable. Wimberly is right in thinking that it is no defence of propaganda to say that it is an attempt at social control in the service of securing the "correct” outcomes. But it is hard for me to see how any attempt to “democratize subjectification”—to give people power over the sorts of people they will be, and to take that power away from the propagandists—could, even if successful, deal with Ellul’s grounds for pessimism, which go far, far deeper.

I’m trying to avoid the academic’s tendency to hedge and qualify in this post. But I want to make one thing clear from the outset. While some people use the idea in annoying ways, epistemic trespassing is clearly A Real Thing. First, while people may abuse the status it bestows, expertise is real: some people genuinely have knowledge and skills that others lack. Second, one way in which experts abuse the status is precisely by leveraging it to give unearned weight to their pronouncements on matters outside their area of expertise. But that’s what epistemic trespassing is. The question I want to answer in this post is what is wrong with it.

The Division of Cognitive Labour

While he didn’t invent the idea of solving a complex problem by dividing up tasks, the philosopher of science Philip Kitcher is credited with developing an important account of the division of cognitive labour. His basic insight is that a modern democratic society such as the US or UK is organised so that people take on roles that match their competences and skills. Meteorologists study weather patterns so we don’t need to rely on simple intuition or experience to predict the weather. Epidemiologists study disease transmission so we have a way of predicting the spread of a new disease. These experts use their specialised knowledge to solve the problems that we, as a society, need solving.

Obviously, this is an idealisation. Kitcher wasn’t blind to the problems with our actual divisions of cognitive labour or our actual societies. His point was rather that, to the extent that complex societies manage to solve the problems they face, they typically do so as a result of a functioning division of cognitive labour. The division of labour works well when the conditions are right. We get the benefits of specialisation—deep expertise in particular domains—combined with the benefits of coordination—different areas of expertise contributing to our overall knowledge and to solving many of the problems we face. But whether we get these benefits depends on whether the division or system is working as it should.

Why Epistemic Trespassing Is Inevitable

Any division of cognitive labour will break down if people don’t stick to their assigned roles. If meteorologists start trying to be epidemiologists, the whole thing will come crashing down. But blatant examples of epistemic trespassing—a physicist moonlighting as a philosopher, a psychologist moonlighting as an international relations analyst—are usually fairly easy to spot. They are less insidious than more complex cases of epistemic trespassing.

One kind of complex case occurs when an expert in some area combines their expertise in that area with a set of values that yield a recommendation about what we should do. This is what our epidemiologist is doing. They have genuine expertise in the distribution, patterns, and causes of diseases and health conditions in a population. What they probably don’t have is expertise in how to weigh up the potential health costs of a virus becoming widely spread within a population against the economic costs of shutting that society down. When they confidently proclaim that the scientific case for lockdowns is clear, they pretend to a kind of expertise that they most likely lack.

Another kind of complex case occurs when an expert in one area can make a credible case that their expertise extends into another, related area. A psychologist having views about the causes of mental health problems in teenagers is not that surprising. It may be that some psychologists have views about this that reflect their genuine expertise in multiple different but related areas. But it is not hard to call to mind cases where someone really does seem to be “reaching”. It’s not that what they have to say about mental health has no value. It’s just that we shouldn’t view their pronouncements about it as invested with the sort of authority that expertise typically bestows. It’s a conversation starter, an idea to consider. It’s not something that should be taken as remotely authoritative.

But remember: I don’t want to psychologise the epistemic trespasser. I want to understand the problems with our systems for dividing cognitive labour that lead to epistemic trespassing. Let me highlight two sets of problems.

The first set of problems concerns the venues in which experts communicate with the public. Most of these venues operate with quite severe time and format constraints. With the exception of the long form podcast, interviews with experts, whether online, on TV or in print, tend to be brief and to the point. There is little time or space for the expert to qualify their assertions or their authority to make those assertions. There is often little incentive for the expert to make any qualifications at all. Complex trade-offs and uncertainty don't make for compelling viewing or listening, even when they're the most accurate representation of the current state of knowledge.

The second set concerns the nexus between experts, policy makers and politicians. One of the features of a technocratic society such as ours is that elected officials typically present themselves as deferring to “expert opinion” on complex matters, such as whether to instigate a national lockdown in the face of a new virus. Our politicians often find this convenient: they can outsource difficult decisions to scientists rather than taking responsibility for their own choices. For someone faced with the decision whether to order a nationwide lockdown, it's politically expedient to defer to a scientist who has a model showing that only a lockdown will avoid millions of deaths. If the decision proves right, the politician can claim credit; if it proves wrong, they can say they were simply "following the science."

This environment creates some problematic political dynamics. One of them is what Jeffrey Friedman calls the "spiral of conviction"—a system that systematically selects for experts willing to exceed their competence. Media, politicians and policymakers often don't want careful, qualified advice that acknowledges uncertainty; they want definitive recommendations that can justify their decisions and provide political cover. This means that experts who are willing to speak beyond their expertise—who can provide the clear, confident answers that politicians need or the public want—are more likely to be consulted, quoted, and elevated to positions of influence. An epidemiologist who says "the data on school closures is mixed and involves complex trade-offs" is less useful to a politician than one who says "schools must close to save lives” ("or “there is no reason at all to close schools”). Meanwhile, experts who carefully qualify their statements or acknowledge the limits of their knowledge often find themselves marginalised in policy debates, or in public discussions more generally.

These features of our media and political environment can put experts in difficult positions, where the roles they're asked to play don't match well with the division of cognitive labour as ideally conceived. Of course, some experts might decide to “play the system”, exacerbating all of the problems I have discussed. But to my mind the more important point is that these problems exist irrespective of the motivations of individual experts. This is why, in my view, the problem of epistemic trespassing is inevitable in a society like ours. It doesn’t exist because our expert class are particularly power hungry and venal. It exists because we have a system that incentivises and rewards epistemic trespassing.

How Politicisation Makes Things Worse

I’ve already explained why I think epistemic trespassing is inevitable. There’s another political dynamic that makes things even worse. Many have commented on the increasingly politicised nature of scientific and other “knowledge” institutions. I have written about this dynamic before and the point I want to make about it is quite subtle. The main reason why politicisation is often a bad thing is that it produces a particularly destructive dynamic. Once something, like an institution or a view on some scientific issue, becomes politicised, it is very hard to depoliticise it. If you don’t like the political valence it has acquired, your only options are surrender, or to fight a political battle against it. If you don’t mind the political valence but would rather it didn’t attach to the institution or view, it is hard for your defences of the view to not be interpreted as moves in the political battle by your opponents. (Presumably some people wanted it to be politicised in the first place, so they’ll be happy).

We can apply this to politicised science. Universities, and knowledge-producing institutions have become sites of political battles and, once that happens, it is very hard to defuse the situation. Backing down means letting the "other side" win. There are, of course, those who want to take a more high-minded approach, equivalent to nuclear disarmament. But the campaign for nuclear disarmament has not been very successful. Just as the unilateral disarmer leaves themselves vulnerable to those who have not disarmed, the researcher or institution that refuses to engage in political battles may find themselves sidelined or irrelevant in increasingly partisan debates.

The upshot is that, in a politicised environment, expertise becomes a resource to be captured rather than something to be respected. Different political factions seek out experts who align with their preferred positions. Experts may feel pressure to align their public statements with their own political commitments, or react to incentives to align their statements with the policy preferences of powerful individuals or groups. This creates a dynamic where the boundaries between expertise and advocacy are blurred simply because people are reacting to the incentives in ways that are broadly instrumentally rational.

The pandemic illustrates many of these points. In many (though not all) countries public health experts found themselves embroiled in heated political battles about individual freedom, masks, lockdowns, school closures and vaccinations. Once they became embroiled in these battles, it was very hard for any individual to avoid their statements about the pandemic becoming more fodder for the political fight. A careful, qualified statement about mask effectiveness or the weak evidence basis for school closures might be seen as giving ammunition to those who wanted to remove mask mandates or re-open all the schools. A more definitive statement might be taken as justifying those who wanted to keep—or even tighten—existing restrictions. This makes it difficult to engage in public health communication in a way that is not strategic and political. If what you say will be interpreted as a move in a “political game” regardless of what you do, why not play the game yourself?

A Way Forward?

The politicisation of expertise is a fundamental challenge to the division of cognitive labour that Kitcher envisions. Indeed, to the extent that Kitcher gives us a good model for how things are meant to work, it is a fundamental challenge to democratic governance. When the division of cognitive labour breaks down—when experts are systematically incentivised to speak beyond their competence—it undermines public trust in expertise and makes the dream of science-informed public policy even harder to achieve. What to do about this?

I’m not big on solutions to complex social and political problems. The style of analysis I favour tends towards explaining why the most important problems are hard to solve, which makes it difficult to really believe in any of the solutions that are typically proffered. To my mind, the most serious problem is the point about political responsibility. I don’t know when it started, or to what extent things were always this way, but our political leaders simply refuse to take responsibility for the decisions that they make. One way in which they do this is by outsourcing these decisions to the experts. But this means that experts are forced into a role that they should not be playing in a democratic society. Unless we can find our political leaders to take more responsibility for their actions, I suspect we are, to put it frankly, fucked.

Happily, I’ve come up with an argument that I like. I’ve thought for a while that Bluesky is the best social media platform. There are some reasons to not like Bluesky. But it has one virtue that most other social media platforms lack: it isn’t addictive. More than that, it is boring. It is so boring that you don’t want to spend much time on it. I deleted the app from my phone months ago and I’ve never considered re-installing it. I simply check it on my browser every so often. In contrast, I’ve deleted and re-installed the X app so often I’ve lost count.

If—as I do—you view social media as of little value but are also—as I am—mildly addicted to it then Bluesky is the best because it is the worst. Imagine a world where the only beer available was Tennent’s lager. (If you’re not Scottish: this is horrible brand of beer). Some people would still drink it because some people are alcoholics (a few also claim to like the taste). But most people would simply give up on drinking beer. Bluesky is the Tennent’s lager of social media sites.

As I said I’m an analytic philosopher. So let me try and and put this argument in the form of premises and a conclusion. Here’s a first pass:

1. A good social media platform minimally disrupts your actual life.

2. The main way in which social media platforms disrupt your life is by being addictive.

3. Bluesky does not disrupt your life because it is the opposite of addictive: it is actively boring.

4. Bluesky is a good social media platform.

Let me say a word or two in support of each premise. Premise 3 is widely accepted; the only people who seem to like Bluesky are a small group of very active users. Everyone else is increasingly bored of it. So we should focus on premises 1 and 2, which is where the action is.

There are many reasons why people are worried about social media: it is warping our brains, it is fomenting civil discord, it is a hotbed of fake news and misinformation. Some think it is the main cause of rising rates of mental health problems in young people, particularly teenage girls. But these sorts of reasons have an underlying political valence, typically left wing (worries about social media and mental health might be an exception). The reason I offer for being worried about social media is more ecumenical and should enjoy multi-partisan support: it sucks up huge amounts of our time and attention. The main way in which it does this is by being incredibly addictive.

There are however two problem with this first pass version of the argument. The first is that it only shows that Bluesky is a good social media platform, not that it is the best one. There may be another social media platform that is even better than Bluesky by the metric of being extremely boring. The second problem is that, if all that can be said about Bluesky is that it is boring, it is unclear why anyone uses it, or should use it. But for a platform to be the best social media platform it is surely a requirement that people actually have a reason to use it.

There is a better version of the argument that can avoid both objections. It starts by noting that there are reasons why someone might use a social media platform even if it is boring. Take Facebook, for example, which has been a car crash for as long as anyone can remember. Why do people still use it? I suspect the main reason is a combination of inertia and network effects; closing an old Facebook account would mean finding a new way of keeping in touch with the people you primarily keep in touch with over Facebook. So why isn’t Facebook a better social media platform than Bluesky? Setting aside the question of whether it is quite as tedious, the problem is that for many users the network effects are extremely strong, so much so that they provide a reason to check it fairly regularly. Bluesky is a newer platform, with a user base that is orders of magnitude smaller, and so the network effects are significantly smaller. But they still exist, hence why I still check it from time to time, albeit less regularly than Facebook.

Here then is the better version of the argument:

1. The best social media platform minimally disrupts your actual life while still providing sufficient reason for use.

2. The main way social media platforms disrupt your life is by being addictive, which might be due to the addictive nature of the content on the platform, or the network effects associated with it.

3. Bluesky provides just enough network effects to justify occasional use, but not enough to create addictive checking patterns.

4. No other social media platform achieves such an optimal balance of minimal disruption with sufficient utility.

5. Bluesky is the best social media platform.

This version avoids both problems with the original argument. Premise 4 is, I must admit, a hostage to fortune: another social media platform may come along that strikes an even more optimal balance than Bluesky. But, until that happens, Bluesky is the best social media platform.

First, a confession: I’m a humanities academic, but I’m really bad at reading books that have anything to do with my research, or non-fiction in general. I far prefer fiction. (I am the mythical man who prefers fiction to non-fiction). I’ve been on research leave since January, and I started out with grand plans to read one “work” book a week. It’s now June and I’ve only managed 11. They are:

1. Jacques Ellul, Propaganda: The Formation of Men’s Attitudes

2. Jeffrey Friedman, Power without Knowledge: A Critique of Technocracy

3. Michael Hannon and Elise Woodard, Political Epistemology

4. Jonathan Ichikawa, Epistemic Courage

5. Walter Lippmann, Public Opinion

6. Randal Marlin, Propaganda and the Ethics of Persuasion

7. Fabien Medvecky and Joan Leach, An Ethics of Science Communication

8. Tristan J. Rodgers, Conservatism: Past and Present

9. Blake Roeber, Political Humility: The Limits of Knowledge in Our Partisan Political Climate

10. Jacob Russell and Dennis Patterson, The Weaponization of Expertise: How Elites Fuel Populism

11. Justin Tosi and Brandon Warmke, Why It’s OK to Mind Your Own Business

What matters is not how much you read but what you gain from that reading. Or so I tell myself. I enjoyed some of these books so much I wrote about them on here. Ellul offers an insightful analysis of propaganda in modern democratic societies, on which we can view propaganda as a set of techniques for social control (I found this analysis so insightful that I wrote about it again). Friedman develops a provocative critique of technocracy that is, as some like to say, “directionally correct” (as in: he has a point but he goes way too far). Ichikawa has some interesting things to say about the connection between scepticism and conservatism, though I don’t agree with his claim that a sceptic must be some sort of Burkean conservative. Roeber has an argument for a kind of scepticism about political knowledge that I myself defend, though I like my argument better. Russell and Patterson make a lot of good points about the pathologies of the expert class, but have surprisingly little to say about the ways in which certain members of that class leverage public distrust of experts to their own advantage. Finally, I liked Tosi and Warmke’s book because I enjoy the quiet life and I’m very open to someone who wants to provide a philosophical justification for my existing preferences.

I also enjoyed some of the books I didn’t write about. Hannon and Woodward’s overview of political epistemology is about as fun as a relatively accessible and comprehsive overview of a field of academic research can be (you can buy it here). Rodgers’ book on conservatism helped me figure out my own relationship with certain strands in conservative thought to which I am sympathetic (it is also an attempt, albeit in my view quite unsuccessful, to provide a degree of philosophical support for contemporary right wing populism). Lippmann is full of insights, many of which seem under-appreciated in my own field of research, political epistemology (I draw on some of them in my pieces on Ellul and Friedman). You may notice I haven’t mentioned two of the books. I leave the reader to draw the obvious implicature.

That’s the non-fiction. I managed to read a lot more fiction. In case it’s of interest to anyone, here’s the list:

1. Pat Barker, Regeneration

2. Heinrich Böll, The Silent Angel

3. Heinrich Böll, Billiards at Half-Past Nine

4. Heinrich Böll, And Where Were You, Adam?

5. Charlotte Brontë, Jane Eyre

6. Emily Brontë, Wuthering Heights

7. Adolfo Bioy Casares, The Invention of Morel

8. Graeme Macrae Burnet, His Bloody Project

9. William Golding, The Lord of the Flies

10. Graham Greene, The End of the Affair

11. Andrey Kurkov, Silver Bone

12. Nella Larsen, Quicksand/Passing

13. Sinclair Lewis, It Can’t Happen Here (not the best, but surprisingly insightful)

14. Cormac McCarthy, The Crossing (re-read)

15. Ian McEwan, On Chesil Beach

16. William McIlvanney, Laidlaw

17. William McIlvanney, The Papers of Tony Veitch

18. Herman Melville, Bartleby the Scrivener, Benito Cereno, Billy Budd, a few others

19. Juan Rulfo, Pedro Páramo

20. Nan Shephard, The Quarry Wood

21. Elizabeth Strout, Tell Me Everything

22. Olga Togarczuk, Flights

23. Yevgeny Zamyatin, We (I wrote about this in connection with Ellul)

I’m not a particularly discriminating reader. I tend to find at least some value in everything I read (I make two exceptions: Portnoy’s Complaint and Portrait of an Artist as a Young Man. Perhaps I’ll explain why in a later post). But I can’t say I loved all of these. I found Lord of the Flies remarkably dull, though I appreciate what Golding was trying to say about public schools and the peculiarly English brand of cruelty they instill. I like Elizabeth Strout—Olive Kitteridge is a fantastic book—but her Lucy Barton series (of which this is a part) is increasingly verging into elaborate self-parody. Finally—and some may view this as sacrilege—I learned that Melville can often be quite dull (Bartleby is brilliant though).

My favourites were Barker, Böll, the Brontës and Shephard. Barker’s Regeneration is the (largely fictional) story of Siegfried Sassoon’s stay at a psychiatric institution for shell-shocked officers during WW1. Böll was a revelation. I started with The Silent Angel, which wasn’t published in his lifetime (frankly, you can see why). But Billiards at Half-Past Nine and And Where Were You, Adam? are something else entirely. Billiards is set after WW2 (apart from some messing with timelines), And Where Were You, Adam? is set right at the end. They are about the impossibility of reconciling oneself to what Germany had done (Billiards) and the senselessness of what it did (Adam). It would be trite to describe either of them, or Regeneration, as “anti-war” books. They take the fact that war—any war, but particularly the war waged by the Nazis—is a moral outrage and ask how one can live in the face of moral outrage. The answer is somewhere between “you can’t” (take the grandmother in Billiards who colludes in the pretence that the loss of much of her family drove her insane) and “with enormous difficulty” (her son, whose response is to live a life devoid of any spontaneity, or Sassoon in Regeneration).

Shephard’s The Quarry Wood is often compared with Lewis Grassic Gibbon’s Sunset Song. The comparison is, in some ways, apt. Both are stories of young women in the northeast of Scotland, caught between their family obligations and their developing senses of self. While I don’t recommend reading fiction to learn about the present (there are better reasons), you can’t read something like Quarry Wood without thinking of the refrain, popular in some quarters, that the solution to the problems of modernity—atomisation, social isolation, a lost sense of meaning and purpose—is to return to an imagined pre-modernity, where you did not choose a set of social roles but had those roles thrust upon you by accident of birth. To be sure, Martha Ironside (the protagonist) doesn’t struggle with a lack of purpose. Her problem is that she lacks the freedom to decide on that purpose—or anything else—for herself. She yearns for the life of the mind; what she gets is the chance to look after assorted relatives, some of whom at least have the decency to express some gratitude for it. Whatever the problems of modernity, I suspect that most would choose them over the total lack of freedom that Martha faces, or indeed that the vast majority of people who have ever lived have faced.

I will credit the reader with knowing enough about Jane Eyre and Wuthering Heights not to need a brief summary. I had always resisted reading the Brontës out of a vague suspicion of English literature (emphasis on the English) that I can’t exactly justify. The suspicion is something like this: the classic English novel is too mannered, too conventional, too neat and tidy. It may criticise social conventions, but it will do so in a restrained way, skewering while still somehow upholding what it is nominally attacking (yes, I’m trying to describe Austen). It may be very clever. It may, like Middlemarch, even approach genius. But it will always be missing something—the madness of many Russian novels (I don’t just mean Dostoyevsky), the fine sense of the absurd present in many great works of European literature, the humour of a Flann O’Brien. I wouldn’t entirely exempt Jane Eyre from these charges; it is, above all else, a deeply moralistic book, if also a very strange one. But Wuthering Heights is something else entirely. I thought it was wonderful.

When I started this project, I expected to focus on what the philosophers had to say about propaganda. I am a philosopher, not a social theorist. But I’ve found that there is a lot of value in these books—perhaps more than in the contemporary philosophical literature on propaganda. This post is an attempt to explain why.

When philosophers and social theorists talk about propaganda, they often seem to be talking about different things. The philosopher wants a definition of propaganda—a set of conditions that cleanly separate propaganda from other forms and techniques of communication. Moreover, the philosopher typically wants a definition that fits with our intuitions about what propaganda is, particularly the intuition that propaganda is a bad thing.1 The social theorist, in contrast, treats propaganda as a sociological phenomenon, a feature of modern societies which has certain functions and serves certain purposes. Where the philosopher wants to know what propaganda is, the sociologist wants to know what propaganda does.

It might seem like these approaches should be compatible, indeed quite separate. You might agree on what counts as propaganda while disagreeing about its social role, or accept a particular sociological account while disputing the proper definition. But any attempt to maintain this independence faces some serious problems. Some definitions might smuggle in sociological assumptions. A sociological analysis might resist the sort of conceptual clarity and precision that philosophers want. The result is that these two approaches, the philosophical and the sociological, sometimes pull in quite different directions.

The Sociology of Propaganda

Let me start by examining three figures—Edward Bernays, Jacques Ellul and Walter Lippmann—who are good examples of how a sociologist might think about propaganda. Their aim is to understand propaganda as a social system. Put together, they tell something approaching a complete story: Lippmann diagnoses a central epistemological problem, Bernays offers a technological solution, and Ellul highlights the dystopian nature of that solution.

In Public Opinion Lippmann frames one of the most basic problems of political epistemology: how do we form an opinion about the world outside our own immediate experience? As he put it:

"In putting together our public opinions, not only do we have to picture more space than we can see with our eyes, and more time than we can feel, but we have to describe and judge more people, more actions, more things than we can ever count, or vividly imagine"

Lippmann’s answer to this problem is that we rely on stereotypes—interpretive frames that we use to filter and organize the overwhelming mass of information confronting us. But, while we can’t do anything but use them, these interpretive frames inevitably distort our understanding. They are simplifying schemes that impose order but distort our picture of the world. The result is that the world, or at least large portions of it, frequently defies our attempts to understand it.

You might think that someone who endorses this view should also endorse a different view: people will tend to have different, sometimes radically different, views about the world because they are using very different frames and stereotypes to interpret it. But, at least in Public Opinion, Lippmann does not endorse this view, or even defend the value of “ideational heterogeneity” (people having lots of different ideas about the world). Instead, he argues that democratic governance requires a high degree of convergence in public opinion. Any democratic society is going to need a tool to manufacture consent:

"That the manufacture of consent is capable of great refinements no one, I think, denies. The process by which public opinions arise is certainly no less intricate than it has appeared in these pages, and the opportunities for manipulation open to anyone who understands the process are plain enough. ... [A]s a result of psychological research, coupled with the modern means of communication, the practice of democracy has turned a corner. A revolution is taking place, infinitely more significant than any shifting of economic power.... Under the impact of propaganda, not necessarily in the sinister meaning of the word alone, the old constants of our thinking have become variables."

Lippmann’s view is something like this: in a democratic society, certain ideas need to be widely agreed upon because, otherwise, there would not be sufficient public support for the things that the government might want to do. Propaganda is a tool for manufacturing this consent.

While Lippmann recognised the necessity for propaganda, he doesn’t tell us that much about propaganda or how the propagandist operates. In Propaganda Bernays is more expansive. But, where Lippmann recognise the necessity for propaganda as a means of manufacturing consent, Bernays is positively enthusiastic about it. I’ll let Bernays speak for himself:

“In theory, every citizen makes up his mind on public questions and matters of private conduct. In practice, if all men had to study for themselves the abstruse economic, political and ethical data involved in every question, they would find it impossible to come to a conclusion without anything. We have voluntarily agreed to let an invisible government sift the data and high-spot the outstanding issue so that our field of choice shall be narrowed to practical proportions” (p. 38).

“Universal literacy was supposed to educate the common man to control his environment … But instead of a mind, universal literacy has given him rubber stamps, rubber stamps inked with advertising slogans, with editorials, with published scientific data, with the trivialities of the tabloids and the platitudes of history, but quite innocent of original thought. Each man’s rubber stamps are the duplicates of millions of others, so that when those millions are exposed to the same stimuli, all received identical imprints. It may seem an exaggeration to say that the American public gets most of its ideas in this wholesale fashion. The mechanism by which ideas are disseminated on a large scale is propaganda, in the broad sense of an organized effort to spread a particular belief or doctrine.” (p. 48).

Propaganda is a chilling book. It is full of brash pronouncements about the power of propaganda to shape attitudes and behaviour, to move people to act and think in the ways desired by the propagandist. Even if you are—as I certainly am—inclined to view most of these claims as hyperbole, perhaps even themselves propaganda, what is chilling is not Bernays’ confidence in the power of propaganda but his conviction that it can be a good thing.

For Bernays, the difference between good and bad propaganda is simply that good propaganda is in the service of truth whereas bad propaganda is not. It isn’t hard to see the problem here. The propagandist may themselves be a “true believer”, or—more often---acting on behalf of someone who is a true believer. They are therefore either in no position to recognise that they are propounding bad propaganda, or have a vested interest in not recognising that this is the case. Their audience is in an even worse position; for Bernays, and for Lippmann, it is propaganda which shapes the audience’s opinions in the first place, so someone subjected to a constant stream of bad propaganda is going to be in no position to see that this is so.

What's striking about Bernays' account is not just his enthusiasm for propaganda, but his focus on propaganda as a comprehensive social technology. Bernays views propaganda as a system for managing public opinion and he has a vision where a technocratic elite control this system. Ellul views propaganda in the same way but he is—to say the least—less excited about it. For Ellul, propaganda is "the air we breathe"—it is a necessary and ubiquitous feature of modern technological society. Modern mass society creates problems that only propaganda can solve: how do we integrate millions of people into complex systems they cannot hope to understand? How do you coordinate behaviour across a vast population? How do you ensure that members of a society have enough in common for that society to function?

Propaganda is, for Ellul, the social glue that holds mass society together. But Ellul is less than sanguine about mass society, or about our prospects for escaping propaganda. Indeed, the very tools we might use to try and resist it—education, rational analysis, democratic participation—are themselves part of the system shaped by propaganda. It is hard to see how we could step outside the system to critique it when our critiques are formed within it.

One important feature of Ellul’s account of propaganda is that it is as much a feature of democratic political systems as it is of authoritarian or fascist ones. Propaganda always serves to uphold the system, but different systems—as well as different political purposes and causes—call for different kinds or techniques of propaganda. For Ellul, democratic political systems require more sophisticated propaganda techniques. A democratic state requires popular participation and support. But public opinion is slow to form, unstable, and unpredictable. Propaganda is the means by which democratic governments “channel and shape” public opinion in ways that support their political ends.

A crucial respect in which propaganda in democratic states is more sophisticated is that it often takes the form of rational persuasion.

“There is such a thing as rational propaganda, just as there is rational advertising. Advertisements for automobiles or electrical appliances are generally based on technical descriptions or proven performance—rational elements used for advertising purposes. Similarly, there is a propaganda based exclusively on facts, statistics, economic ideas.”

In fact, Ellul thinks that, as society progresses, propaganda becomes increasingly fact-based:

“the more progress we make, the more propaganda becomes rational and the more it is based on serious arguments, on dissemination of knowledge, on factual information, figures, and statistics.”

It is simply a mistake to view propaganda as necessarily connected with falsehood, lies, and deception. Propaganda in the form of blatant falsehoods is simply a symptom of insufficiently advanced society. The more developed a society, the more likely it is to utilise rational propaganda. To be sure, the (rational) propagandist does not simply provide facts, figures and statistics. They provide a sample of relevant facts designed to support the opinions they would like their audience to form and reinforce the behaviours they would like that audience to exhibit.

Here’s an example of how this works in practice. A crude propaganda campaign would simply declare, whether in simple words or crude images, that immigrants are dangerous. A more sophisticated, more rational campaign would present crime statistics, unemployment figures, and the like—all real, all with sources—but carefully selected and framed in a way that leads the audience towards the same conclusion as the crude propaganda campaign. (If you want an example with a different political valence: imagine a public vaccination campaign that cites real figures about the efficacy of the vaccine but neglects to break those figures down by demographic group so as to prevent the audience from considering whether those figures are directly relevant to them).

Ellul offers what appears to be a definition of propaganda:

"Propaganda is a set of methods employed by an organized group that wants to bring about the active or passive participation in its actions of a mass of individuals, psychologically unified through psychological manipulation and incorporated in an organization."

But this is a very open-ended definition. It essentially defines propaganda as a set of methods of social control used by a group (such as a political party) on a population that has been “primed” to be receptive to these methods. This priming is itself typically the result of previous propaganda operations. If you are worried that this sounds a bit circular, bear in mind that Ellul has a typology of propaganda methods, starting from the most basic ones designed to move already existing groups of people to action, often in the form of civil disorder and unrest. The circularity is therefore not a flaw in Ellul’s analysis. Propaganda creates the conditions for more propaganda. It shapes people who then demand the kind of simplified, emotionally satisfying information that propaganda provides. The target of propaganda becomes complicit in their own manipulation.

The Philosophy of Propaganda

Where a social theorist likes Ellul wants to understand what propaganda does—how it works, why it exists—the philosopher tends to want to try and define propaganda. Typically—though with some exceptions2—the definition is something like this: propaganda is “epistemically flawed” messaging. The propagandist crafts messages that lead people to form beliefs that are false, or at least inapt or misleading in some way.

Sheryl Tuttle Ross's influential account of propaganda is illustrative. Ross sets out the capture the “pejorative” character of propaganda: propaganda is a form of persuasion that is in some way illegitimate. Here is her definition:

Propaganda is (1) an epistemically defective message (2) used with the intention to persuade (3) a socially significant group of people (4) on behalf of a political institution, organization, or cause.

The key conditions here is (1), which captures the pejorative character of propaganda: propaganda is bad because the propagandist produces messages that are epistemically defective. A message can be epistemically defective simply because it is false. But it can be defective in subtler ways. For example, a video depicting an immigrant committing a crime may be an accurate representation of something that happened. But it invites the viewer to form false beliefs about immigrant criminality in general.

Other philosophers have built on Ross’s definition. Constant Bonard, Filippo Contesi and Teresa Marques have an account which replaces messages with the broader category of “communicative acts” (commands, questions, pictures, gestures). They also have a more expansive conception of what can make a communicative act/message epistemically defective, and they specify that propaganda must be produced for political ends (a state education system is, for them, not inherently propagandistic).

Amelia Godber and Gloria Origgi focus less on the falsity of what is communicated by propaganda and more on the ways in which propaganda interferes with our ability to rationally deliberate. For them, the key feature of propaganda is not so much what it tries to get us to believe but how it tries to get us to believe it. The propagandist may present us with facts, evidence, reasons. But they do so in ways that distort our ability to think about those facts. They may present a very biased subset of the relevant facts. Or they may present a more representative subset, but in ways that impede our ability to properly assess those facts.

Beyond the similarity in the definitions of propaganda they offer, these accounts have a few other things in common. First, they are all attempts to capture a “commonsensical” or “pretheoretical” conception of propaganda. They are meant to accord with our “intuitions” about what is (and what isn’t) propaganda. Thus, because propaganda is typically viewed as a bad thing, any definition must capture this “badness”. Further, despite the fact that there are deep historical connections between propaganda, marketing and public relations, these connections are not part of the “ordinary” concept, so we don’t want a philosophical account of propaganda to reflect these connections.

Second, they all view propaganda as primarily influencing beliefs. The propagandist wants to get their target audience to believe things—that immigrants are criminals, that the enemy is committing atrocities, that their country is a force for good. To be sure, the reason why the propagandist wants to get their target to believe things is often as a means to influence their behaviour. Someone who thinks the enemy is evil is going to be more inclined to go to war to fight the enemy. But the influence on behaviour is indirect, via influence on belief.

Third, while they view propaganda as inherently a communicative act—the propagandist is communicating something to their target audience—they don’t have anything to say about the ways in which the propagandist might try to shape that audience, or about the sort of relationship between propagandist and audience that is required for propaganda to achieve any sort of uptake. The idea seems to be that we can think about propaganda purely in terms of the features of individual acts of propaganda. This atomistic account stands in stark contrast to more holistic accounts, which emphasise the broader communicative ecology—how the propagandist cultivates their audience over time, and the institutional frameworks that help them do this.

But what is wrong with an atomistic, belief-centric account of propaganda? What is wrong with trying to make sense of ordinary intuitions? Who could object to someone trying to give analytical clarity to a messy concept like propaganda?

Ellul versus the Philosophers

Maybe I’m just not a very good philosopher, but I find what Ellul—and for that matter Bernays—has to say about propaganda to be a good deal more insightful than what most philosophers have to say about it. Some of this is perhaps just a matter of taste; I’m not convinced that there is an ordinary concept of propaganda to analyse, or that the attempt to analyse it is much more than a way of making a particular way of thinking about propaganda—a way that might carry an implicit political valence—more precise. We learn how a 21st-century philosopher understands propaganda, not what propaganda is. Ellul’s approach is, I think, preferable on methodological grounds: it simply gives us more insight into the phenomenon we are meant to be interested in.

Ellul does of course have a definition of propaganda: it is a set of methods that an organized group employs in order to bring it about that a mass of individuals participates, whether actively or passively, in its actions. But this isn’t meant to be the sort of definition that clarifies that much (or the sort that is intended to provide necessary and sufficient conditions for something being an act of propaganda). The methods that the propagandist employs may vary immensely: the propagandist charged with creating civil unrest needs one set of tactics (agitation propaganda); the propagandist charged with ensuring the stable functioning of a mature democratic society needs a quite different set (Ellul calls this “integration propaganda”). What unifies these disparate methods is that they are means of controlling a population. Thus a simpler formulation of Ellul’s account of propaganda would simply be that it is a set of techniques for achieving social control. This tells us the sort of thing propaganda is, but it doesn’t define it in any meaningful sense.

Where Ellul offers real insight is when it comes to how propaganda (when it works) manages to achieve a degree of social control. But Ellul’s explanation of this requires rejecting two of the features of the philosophers’ accounts of propaganda. First, for Ellul, the propagandist aims at shaping action rather than belief:

The aim of modern propaganda is no longer to modify ideas but to provoke action. It is no longer to change adherence to a doctrine but to make the individual cling irrationally to a course of action.

This passage is a little bit misleading, in that Ellul did think that propaganda also worked at the level of ideas. But it typically works not by modifying ideas but by giving existing ideas—and prejudices—a firmer shape. Ellul called this “psychological crystallisation”:

Propaganda furnishes objectives, organizes the traits of an individual's personality into a system, and freezes them into a mold... prejudices that exist about any event become greatly reinforced and hardened by propaganda; the individual is told that he is right in harboring them; he discovers reasons and justifications for a prejudice when it is clearly shared by many and proclaimed openly.

The result is the transmutation of existing ideas and prejudices into firmer judgements:

he [the target of propaganda] now has a supply of ready-made judgments where he had only some vague notions before the propaganda set in; and those judgments permit him to face any situation.

As these passages should make clear, Ellul isn’t really talking about propaganda as a means of shaping and modifying beliefs in the philosopher’s sense of the term. It would be more accurate to say that, for Ellul, the aim of propaganda is to shape its target—to make them the sort of person who will do (and want to do) what the propagandist would like them to do. This will inevitably lead to the target of propaganda having certain beliefs, but this is, if you like, simply an upshot of the larger “game” that the propagandist is engaged in.

Second, Ellul doesn’t think it makes much sense to think about propaganda in terms of isolated, one-off acts of propaganda. In fact, he thinks that, if you view propaganda this way, you will come to the conclusion that it is singularly ineffective (as above: the propagandist needs to start by creating a willing audience for their propaganda). Ellul takes the opposite approach: propaganda is not an aberration, a form of defective communication that individuals and state actors engage in from time to time. It is everywhere. It is what holds modern society together—or, occasionally, what you might use to fracture a society, as when propaganda is used in the service of starting a revolution.

Of course, there is no such thing as “modern society”. There are multiple modern societies, and different societies work in very different ways. Ellul doesn’t think there is a single set of techniques for keeping any modern society together. Rather he thinks that, in any society, there will be forms of propaganda that help hold that society together. In cohesive democratic societies integration propaganda is particularly important. Integration propaganda operates through advertising, education, and entertainment, creating conformity not through force but through the subtle presentation of social norms. For instance, the promotion of the "American Dream" serves as a form of integration propaganda that shapes values and behaviours without appearing to coerce anyone into accepting those values or performing those behaviours.

For Ellul, a technological society isn't merely one with advanced gadgets; it's a society organised around technique, efficiency, and instrumental rationality. Any technological society has advanced physical infrastructure—roads, hospitals, schools, airports, etc. Propaganda is the counterpart to this physical infrastructure—it is a means of ensuring compliance from the members of a modern technological society. Compliance requires more than just simply “going along” with things. It requires psychological integration. For Ellul, propaganda always has this totalising nature—though in authoritarian regimes the totalising nature takes on an undeniably more frightening aspect than in democratic regimes.

You might wonder whether Ellul ever bothers to give any evidence that propaganda—understood in this distinctive way—actually works. Let me be clear: he is a social theorist in the grand tradition. Do not read Propaganda expecting to find a wide array of citations to studies that provide bits of empirical evidence for the various claims he advances. But in a way to ask the question is to misunderstand what Ellul is trying to do. His account of propaganda is intended as an explanation of something that he thinks is undeniable and is very much in need of explanation: what holds a complex modern society together? His answer is that modern societies are held together—to the extent that they are held together—by a complex set of techniques for social control. That is, they are held together by propaganda.

Ellul’s sociological approach to propaganda might frustrate the philosopher looking for a nice, clean set of conditions for counting something as propaganda. But, whatever its limitations, it captures some of the essential—or at least the most interesting—features of propaganda that the philosophers’ approach is liable to miss. By fixating on propaganda as a form of messaging that implants false or misleading beliefs, we miss the more important impact of character on action and character. By treating propaganda as a discrete category of defective communication, or by focusing on one-off acts of propaganda, we fail to understand propaganda as a pervasive feature of technological society. Finally, by viewing propaganda as a central means by which states try to control their citizens we come to understand why, though many take such a dim view of propaganda, it is everywhere. We may want to do away with propaganda. But, if Ellul is right, that might require also doing away with modern society. Whatever the problems of modernity, this seems like a high price to pay. We are therefore left in the uneasy position of both recognising the necessity of certain forms of propaganda and wishing we could live without it.

Ichikawa draws a connection between scepticism and conservatism in order to criticise scepticism: the sceptic is not meant to gain by the comparison. While I'm going to try and tease his argument apart because I want to resist the comparison, this is not because I see no value in conservative political thought. I find much to agree with in the conservative ideas of Edmund Burke and David Hume: the limitations and imperfections of reason, of abstract political principles, and of human beings. In particular, I recognise myself in Michael Oakeshott's much quoted articulation of the "conservative disposition" which is:

to prefer the familiar to the unknown, to prefer the tried to the untried, fact to mystery, the actual to the possible, the limited to the unbounded, the near to the distant, the sufficient to the superabundant, the convenient to the perfect, present laughter to utopian bliss.

But, as we will see, what is distinctive of conservative political thought is the fact that these pessimistic and sceptical ideas are combined with something more positive. In Burke, pessimism about human reason is combined with a deep veneration for tradition. In other conservative thinkers, pessimism is combined with religious faith (also present in Burke), natural law, or the spontaneous order of markets. With the exception of markets, which certainly have their uses, I am even more sceptical of these foundations than I am of human reason, if not more so. Conservativism is, if you like, what happens when you only take scepticism so far—far enough to defend your favoured set of conservative ideas, but no farther.

Ichikawa on Scepticism and Conservatism

Here is a brief reconstruction of Ichikawa’s argument (the page references are all to his book):

1. Scepticism creates a bias toward avoiding false beliefs over acquiring true beliefs, which leads people to suspend judgment rather than form beliefs.

2. Since we act based on what we believe, and “a belief that one has no good reason to hold constitutes no good reason to act” (p. 27), suspension of judgement leads to inaction.

3. While inaction preserves the status quo, changing the status quo requires action, which in turn requires beliefs about how to make changes.

4. The sceptic’s preference for inaction embodies the assumption that "things are basically fine the way they are" or at least "tolerable enough to make potential deviations risky" (p. 18).

5. The assumption that the status quo is preferable to uncertain change reflects "the foundational ideas underwriting political conservatism" (p. 29).

6. Therefore, scepticism promotes conservatism by motivating inaction, which "amounts to a tacit endorsement of harmful, even oppressive, features of the status quo" (p. 34).

In summary, Ichikawa argues that scepticism's preference for doubt over belief naturally aligns with conservative politics: when you don't believe strongly enough to act, you default to preserving what already exists. This is political conservatism.

Ichikawa’s starting point is the observation that scepticism embodies a preference for avoiding false beliefs over having true beliefs. The sceptic, says Ichikawa, is someone who is so worried about the risks of getting it wrong that they fail to consider the risks of not getting it right.

This, Ichikawa thinks, means that the sceptic is prone to inaction. The argument here relies on a close tie between belief and action: we act based on what we believe, and our actions are justified only when based on justified beliefs. Because the sceptic has few, if any, justified beliefs, they will rarely have any justification for acting. Indeed, the sceptic may well lack many beliefs at all, in which case it is unclear whether they will have any basis for acting in the first place.

Scepticism functions to preserve the status quo: if you don’t do anything, nothing is going to change. Scepticism therefore has systematic political effects. As Ichikawa puts it:

Doing things is another way of saying making changes to the way things are in the status quo. So skepticism, at least sometimes, will have a tendency to direct one to 'go with the flow.' And this tendency, I think, has a lot to do with some of the foundational ideas underwriting political conservatism (p. 29).

This connection to conservatism is structural, not incidental. Indeed, scepticism is a key component of conservatism:

the negative bias—the idea that it is much more important to make sure we don't believe wrongly than to make sure that we do believe rightly—reflects a particular assumption: that skepticism and suspension of belief are intimately tied up with caution and carefulness. Given plausible connections between epistemology and action, this thought, in turn, is committed to conservative assumptions about the status quo (p. 18).

Of course, “conservatism”, much like “liberalism”, is a contested word. Ichikawa is primarily talking about Burkean conservatism, which I mentioned earlier. This, says Ichikawa, is based on two key ideas: "veneration for tradition as a repository of 'humanity's true wisdom'" and "a cautious approach to social change, based on a deep scepticism about the prospects of attempts to deliberately re-engineer society” (p. 30). The thought is that it is usually wiser to continue an old practice than to run the risk of changing it. Burkean conservatism goes naturally with defending the status quo: the status quo embodies the wisdom of humanity, such as it is, and we should be reluctant to change it—indeed, extremely reluctant if the proposed change is anything approaching radical.

Putting this together, then, we get the following picture of the politics of scepticism. For social change to occur, people need justified beliefs about what changes are needed and how to achieve them. But, as Ichikawa argues, sceptical arguments are intended to show that we don’t have many (or even any) justified beliefs, so scepticism is a recipe for inaction and a justification for the status quo. This, says Ichikawa, is bad:

A bias in favour of the status quo is beneficial to those for whom the status quo is working well, but not to those who suffer under it. So the skeptical tendency will often entrench or perpetuate inequalities or injustices (p. 34).

Ichikawa's point here takes two forms: scepticism can entrench injustice either unintentionally or strategically. On the first form, the idea is that even sincerely held forms of scepticism produce these results—perhaps contrary to the intentions of the sceptic. On the second, the idea is that professions of scepticism are used strategically to serve conservative political ends. As Ichikawa notes, refusing to be convinced of “certain disruptive ideas” can be a good way of preventing action to change the status quo—and of cloaking one’s refusal in the language of critical thinking and rationality. Either way, though, Ichikawa thinks there are close ties between scepticism and political conservatism.

Response 1: The Belief-Action Connection

Ichikawa's argument hinges on a particular theory of human action: we act based on what we believe and, without justified beliefs, we lack justification for action. This, he thinks, creates a direct line from scepticism to inaction (without justified beliefs, you won’t act). From this he concludes that scepticism is allied with conservatism, in that it provides a rationale (or simply cover) for the conservative’s hostility to change. I think his underlying picture of human psychology is overly rationalistic. Scepticism may be a barrier to justified belief, but it is only a barrier to action if you think that action must be based on beliefs, justified or otherwise.

For action to occur, there must be something that spurs you to act. If I’m sitting inside on a chair, something needs to make me get up and go outside. But what provides the spur to action need not be a belief, still less a justified belief. Often, action is more the result of habit than anything else: I get up and go outside because it’s time to go to work and I’m in the habit of going to work. This doesn’t just apply to mundane actions like going to work. It also applies to political action. Or at least I think it applies to my political actions—perhaps I’m simply unusual. When I vote in an election I don’t do so because I have what I take to be a justified belief to the effect that my vote matters or will make any sort of difference. I vote because I have become habituated to do so. When I participate in another futile industrial action by UCU (one of the main unions for employees of universities and colleges in the UK), I don’t (usually) do so because I have what I take to be a justified belief about the value or likelihood of success of the action. I simply view it as the duty of a union member—it is part of what solidarity requires.

Even if you accept Ichikawa’s rationalistic picture of individual human action, it is unclear that we should accept a rationalistic picture of social change and the processes that produce it. However much the conservative may decry it, social change occurs: the decline of the authority of the Church in western Europe, the creation of the welfare state, the loosening of traditional gender roles. It is however a further question why it occurs. On a rationalistic picture like Ichikawa’s, beliefs, justified by the evidence, play an important role in explaining why social change happens: individual people form beliefs about what should be done, and they set about trying to turn those beliefs into reality. If they didn’t have the beliefs, or doubted whether they were justified by the evidence, they wouldn’t have tried to change anything.

It isn’t that this picture is entirely wrong. History furnishes examples of people who had beliefs like this, and who acted on them. These are the leaders with a vision, a strongly held set of beliefs about what the world should be like. We celebrate some of these leaders because we, from our privileged vantage point in the present, recognise the value or truth of the vision, and applaud their success in bringing it into being. We decry, or downright revile, other leaders because of the repugnancy of their vision—we wish Hitler or Stalin had been less convinced in the correctness of their respective views of the world.

But to view social change as largely the result of the actions of ideologues would be to endorse a rationalistic version of the great man theory of history—the view that historical change is primarily driven by the actions of exceptional individuals rather than by broader social, economic, and cultural forces. Social change is typically the result of a huge range of factors, many of which are not particularly amenable to a rationalistic mode of explanation: social conditions (like mass poverty or unemployment), economic forces (like the rise of globalisation in the 20th-century), cultural shifts (like the decline of religious belief in western Europe), contingent events (like a pandemic), technological changes (like the internet, or perhaps AI), and the unpredictable dynamics of social movements themselves. People join social movements for all sorts of reasons: because their friends do, out of a sense of solidarity, because the status quo has simply become intolerable, or because they are swept up in historical events they don’t understand or control.

It is unclear why sceptical ideas would—or for that matter should—stand in the way of these or other forces that drive social change. It is also unclear why they would stand in the way of an individual involving themselves in the social movements that drive social change. Sceptical considerations of the sort I considered above (the difficulty of knowing what the impact of new policies will be) may be a problem for the politician or policy maker devised with solving a social problem like mass unemployment. But they are hardly a reason for someone who is experiencing unemployment to doubt the reality of their situation, or the need that something be done about it. The impetus for social movements typically comes from the conviction that there is a problem—people are starving, they can’t afford a place to live, they are being killed—not any firm belief about how to solve the problem. It would be a grave mistake to view the fact that ordinary people want things like food, safe housing, or an end to violence as entailing that they have any particular view about the right way of achieving these things, or about the political ideology that provides the most useful framework for thinking about them.

The problem then is that Ichikawa’s argument rests on a narrow conception of what might drive social change. The sceptic may think it is hard to have justified beliefs about the best way to reform society or solve particular social problems, but this is only an impediment to political action if you assume that taking action to solve a problem requires having a worked-out solution in advance of trying to solve it. In reality, most social change is the result of trial and error, incremental adjustments, and the attempt to respond to immediate pressures. There is no sceptical argument against attempting to change the world in this manner. What the sceptic wants to say is simply that we shouldn’t expect anyone involved in the attempt to change the world to have arrived at the One Right Way of doing it. Consequently, says the sceptic, we should be wary of allowing anyone to impose their solution on everyone else. Human history provides ample examples to justify this wariness: this is a scepticism based on experience, not abstract philosophical argument.

Response 2: Scepticism as Anti-Political

I do think that scepticism lends itself fairly naturally to a political stance, but that stance isn’t conservative. It is anti-political—a rejection of politics. This stance can, however, easily be confused with conservatism, and in certain situations it may amount to much the same thing.

One way to think of the political valence of scepticism is to view it as the stance you arrive at when you have rejected both radicalism and conservatism. As Ichikawa notes, the right contrast for Burkean conservatism is not liberalism or progressivism but radicalism—the view that existing social arrangements are fundamentally flawed and require comprehensive transformation. While radicalism is not necessarily tied to rationalism (there may be other sources for your ideas about which radical changes are required), it is natural to think of the radical as having the view that we should rethink society based on rational principles. The sceptic is minded to agree with Burke’s criticism of radicalism. It evinces a faith in the ability of humans to rethink social arrangements from rational principles that is at best naive. Societies are too complex for that; attempts to rethink them from the ground up are not just doomed to failure but liable to lead to disaster.

The sceptic parts company with Burke, however, because they think that his conservative alternative relies on equally dubious foundations. Burke’s conservatism doesn't just counsel caution; it makes claims for the wisdom embodied in tradition. It treats inherited institutions—the church, monarchy, aristocracy, common law—as repositories of accumulated wisdom, organically evolved solutions to the problem of social order. Where the radical trusts reason to redesign society, the conservative trusts tradition to preserve it. But why should the sceptic find tradition any more trustworthy than reason?

There are two problems here. The initial problem is just that of providing a defence of tradition. Just as we can view institutions and traditions as embodying the wisdom of the ages, we can also view them as embodying human folly, or simply as the result of a series of historical accidents that occurred for no discernible reason. The mistake here seems similar to the mistake in simplistic applications of the theory of evolution: from the fact that any organism is the product of evolution by natural selection, and so in that sense is adapted to the environment, we cannot conclude that the organism has been designed well, or is functional in any particular respect. We can simply conclude that it functions well enough to have survived.

The second problem is a bit deeper. The problem with appeals to “tradition”—especially when we are talking about the tradition of a country or nation—is that they typically fail to grapple with the fact that traditions are contested. Every complex society contains multiple, often conflicting, traditions: different churches (a problem not unfamiliar to Burke), disputed lines of succession, disputed interpretations of legal tradition, imperial ambition against local autonomy, the countryside against the city, North against South, East against West. The conservative must identify the traditions worth preserving. By their own lights, we cannot do this by appeal to “reason”. What, then, is left? In practice, these disputes are typically resolved when one side “wins”, whether through force, politicking, or historical accident. When the conservative celebrates “our” traditions, they’re often simply venerating the winners of past power struggles.

This is why, when applied consistently, scepticism doesn’t lead to conservatism but to a more radical conclusion: the difficulty of justifying any “big picture” approach to politics. For the sceptic, the conservative’s faith in tradition is as unfounded as the radical’s faith in progress and reason. The consistent sceptic might sound like a conservative when criticising technocratic schemes for social improvement or railing against modernity and the pretense that modern technological society has the cure to humanity’s ills. But, unless they make the mistake of confusing a critique of modernity for a defence of the past, they will not follow the conservative in defending the value of custom and tradition. They will argue for the dangers of any overarching political ideology—including those whose defenders sometimes pretend is not an overarching ideology.

Motivated Scepticism

I said at the outset that I felt compelled to respond to Ichikawa’s argument because, while I am inclined towards scepticism, I don’t think of myself as a political conservative. But I haven’t said anything about the form of scepticism I am inclined towards. Let me finish by saying a bit about it.

One of the points that Ichikawa makes in his discussion is that scepticism can be strategically useful. Imagine you own a coal power station and you want to argue against a politician who is proposing putting a heavy tax on coal production. You don’t have any particular view about whether fossil fuels are causing climate change; you mostly care about preserving your profits. Sceptical arguments are going to be very useful to your attempts to preserve what you view as your interests. You might acknowledge that it looks like fossil fuels are causing rises in global temperatures, but point out that there may be other explanations, or argue that our data is bad, or simply say that we need a lot of evidence before we do things which might make energy more expensive for everyone. If you make these points enough, you’ll most likely end up convincing yourself and suspend judgement on whether fossil fuels are causing climate change. After all, the best way to convince others that you think something is usually to manoeuvre yourself into really thinking it.

There is an important point here: sceptical arguments, if they work, lead us to suspend judgement (that is: not make up our minds) on the issue in question. But, as our example makes clear, you can suspend judgement entirely for reasons of self-interest. It is natural to say that the sceptical arguments you cite as reasons for suspending judgement on whether fossil fuels cause climate change really function as rationalisations. What you want is to suspend judgement; sceptical arguments are simply your means of doing so, and they are an effective means because they allow you to justify suspending both to yourself and others. What I want to do is extend this point: in general, the arguments, evidence and reasons we can cite often function to rationalise positions we have already adopted for reasons that have little to do with the arguments, evidence and reasons we can cite.

This is a familiar phenomenon in politics: someone has a particular set of underlying political commitments—they are a Marxist, a boring liberal, a traditional conservative—and they find a way of making all the evidence they have assembled fit with those political commitments, even when it might look to an outsider like it doesn’t fit at all. There are, broadly speaking, two possible explanations for this happy convergence. The first is that they are simply weighing up the evidence in such a way that it really does support their favoured position. The second is that they are assembling evidence that will fit with their favoured position—and finding ways of explaining away any evidence that does not fit.

Often, the second explanation is more illuminating than the first. That is, I think that it is often the case that the evidence we can cite for our political attitudes are post-hoc rationalisations for those attitudes. We have the attitudes in the first place for very different reasons: they align with our political identities and cultural values, they make us feel happy about ourselves, they fit nicely with what we would like to be true about the world. We then gather the evidence we need to support them.

Importantly, this position does not just apply to our beliefs. It applies to our refusals to believe—to cases where we suspend judgement. This then is a form of scepticism that applies as much to suspension as it does to belief. It is, in some ways, a deeper form of scepticism than familiar forms. It is not about whether we have sufficient evidence for our beliefs, but about the relationship between evidence and belief. On this view, we don’t form beliefs based on evidence and then act accordingly. Instead, we act based on temperament, circumstance and interest, and then marshal evidence to justify what we have already decided. Or, alternatively, we decide not to act based on temperament, circumstance and interest, and the marshal evidence to justify our inaction—to justify our failing to form a belief, our suspending judgement.

This form of scepticism—which I call motivated scepticism—is in some ways more corrosive than the form of scepticism that Ichikawa targets. It suggests that our political attitudes, whether sceptical or dogmatic, conservative or radical, are less the products of careful reasoning than we would like to believe. This is damaging to our self-conceptions.

In other ways it is less threatening. One way in which it is less threatening is that it does not automatically lead to inaction. The whole point is that you should not make the mistake of thinking that suspending judgement is a “safe” stance, a way of avoiding the risks of forming a false belief. Suspending judgement isn’t safe because it might be as much the product of self-interest as belief. Someone—like our owner of the coal power station—who finds themselves citing sceptical arguments to justify their inaction should be suspicious of their reasons for doing so. They should ask themselves whether these arguments simply provide a rationalisation for doing what they would like to do for self-interested reasons.

Of course, someone who is entirely convinced that their course of action is the right one, because it is backed up by a fully justified set of beliefs about the right thing to do, is in much the same position. The upshot of this, though, is that, while scepticism does not provide a neat way of justifying inaction, it does provide a “check” on action. It prompts us to consider whether what we believe to be true, perhaps even obviously true, really is true. The value of motivated scepticism, then, is that it draws out attention to our motivations—whether we’re too quick to believe what we want to be true or strategically suspending judgement. This sort of self-scrutiny may be one of our best safeguards against the twin dangers of dogmatism and convenient doubt.

This narrative involves several inter-related assumptions. First, if people would just recognise genuine expertise, public discourse would improve and societal consensus on important issues would follow. Second, the traditional gatekeepers do (or at least used to do) a good job of figuring things out—finding solutions to complex social problems—and communicating those solutions to the public. Third, one reason why people are unable to recognise genuine expertise is lack of education or faulty critical thinking—people are simply unable to distinguish between reliable and unreliable sources. Finally, experts serve as objective arbiters—they stand above petty partisan political disagreement.

The narrative also lends itself naturally to a proposed solution: we need to increase trust in legitimate experts via better education, improved media literacy, and aggressive content moderation to limit the spread of misinformation. Different proponents of this solution will weigh these components differently: you might prefer to focus on education rather than content moderation, in the hope that better educated subjects won’t be so easily fooled. But all proponents of this solution agree we need to restore deference to credentialed experts and established institutions. They are our primary knowledge authorities, so we need to put our trust in them

That was a bit of a caricature. Anyone—or at least anyone worth engaging with—who seriously defends the common narrative will add a good deal of nuance. Experts and expert institutions aren’t perfect; what, though, is the alternative? But it is a caricature that picks out a recognisable position. We can call it the “technocratic mindset”: experts have the skills and knowledge required to solve complex social problems, so we should give them the job of solving them for us. Our job—unless you happen to be a member of this expert class—is to let them get on with solving them. The problem is that, right now, we—at least, in certain countries—are not letting the experts get on with it. This needs to be remedied, by some combination of education, media literacy, and content moderation.

A recent book, Jacob Russell and Dennis Patterson’s The Weaponization of Expertise, sets out to explain and critique the technocratic mindset and the common narrative about the problem with public discourse that goes with it. For Russell and Patterson, the technocratic mindset is an expression of hubris (experts have the skills and knowledge required to solve all our problems) and the common narrative is completely wrong. It is wrong because it misunderstands both the nature of expertise and the reasons for widespread rejection of it. The solutions put forward by proponents of the common narrative won’t work because they misunderstand the reasons why many people don’t listen to the experts. They mischaracterise rejection of the expert consensus as a simple refusal to listen to what the experts have to say. People know what the experts are saying. They just don’t believe them.

I can’t say that I agree with everything in Russell and Patterson’s book. But those failings are the flip side of the things it does well. They quite rightly skewer assorted members of the expert class for two related failings: they over-estimate their own skills and abilities, and they under-estimate the skills and abilities of ordinary people. As a result, they misunderstand the source of public skepticism about the expert class. But, where Russell and Patterson have a nuanced and sympathetic understanding of these ordinary people, they fail to extend that nuance or sympathy to the experts or the institutions they inhabit. The result sometimes come close to the inverse of the common narrative: the problem isn’t ordinary people but the experts. As a corrective to the pathologies of elite discourse, this is welcome. As a serious attempt to understand our current political moment, it is potentially dangerous in the way that all simplifications are potentially dangerous.

But what, exactly, is their corrective? And why—despite my agreement with many of their central claims—am I not fully on board with it? Let me start with a summary of some of Russell and Paterson's main arguments.

Subscribe now

The Technocratic Mindset

In the first part of the book Russell and Patterson take a closer look at elites. Their core thesis is that elites use claims of expertise to shut down democratic debate on policy questions that involve value judgments. For Russell and Patterson, elites are primarily defined by their educational credentials. These credentials create what they call a "diploma divide" in political and cultural attitudes that drives populism.

This elite class is characterised by what Russell and Patterson call the "technocratic mindset". This mindset "reduces complex social and political questions to technical problems, removing the role of values and public judgment". As they put it:

Public-policy questions are, to them [that is: elite technocrats], ultimately technical puzzles to be solved by experts. This epistemic stance is what interests us about the elite because we think it is this attitude—more than any particular political vision—that ignites the ire of populists.

The COVID-19 pandemic serves as their central case study of expert overreach. During the pandemic, they argue, scientific authority was extended beyond technical matters into domains requiring value judgments. This, they think, was a mistake:

Most troublingly, the idea that we had to defer to science extended even to claims that could not be adjudicated by science alone. Pandemic policies presented trade-offs that could not possibly be resolved without considering domains far beyond fields such as immunology.

Russell and Patterson provide numerous examples of this overreach, but they focus on two pandemic-related policies in particular: school closures and mask mandates. In the case of school closures, they argue that experts advocated for extended closures despite early evidence against their necessity, leading to significant educational harms. Similarly, they argue that mask policies are a case in point of the misuse of expertise during the pandemic: exaggerated claims of efficacy, attempts to stifle debate, and to dismiss any and all contradicting research.

Importantly, Russell and Patterson argue that the problems with expertise are structural and not confined to the pandemic. They cite John Ioannidis's influential paper "Why Most Published Research Findings Are False" to support their claim that methodological limitations, publication bias, and the like, distort scientific research. More generally, they argue that the replication crisis is in large part the result of an academic system that rewards “novel, surprising findings” instead of “careful, nuanced work”. They also argue that researchers deserve some of the blame for the many forms of exaggeration that occur in press releases and media stories about published research.

Populism as Rational Response to Elite Overreach

Russell and Paterson frame populism as a rational response to elite overreach, rather than simple ignorance or resentment. They define populism as:

A reaction to elites running along cultural, economic, and political dimensions and caused by the absence of opportunities for exit and voice.

Their two paradigm examples of populism are the rise of Donald Trump in the US (the book was written before the 2024 election, but their analysis would nicely explain the result) and the 2016 Brexit referendum. But they also understand “science scepticism”, particularly in the context of the COVID-19 pandemic, as a manifestation of the same underlying populist rejection of elite overreach.

For Russell and Paterson, the problem with a lot of discussions of populism is that they are oversimplified. They are oversimplified in a few different ways. One is that they tend to focus on populist politicians and their rhetoric (case in point: Donald Trump), rather than populist voters and their motiviations. Another is that populust voters are typically treated as monolithic movements, when in reality they are diverse coalitions. (Here it is perhaps important to distinguish between the most vocal supporters of a politician and the coalition that votes for them, which in any large electorate is always going to be relatively diverse).

Perhaps the easiest way to see why Russell and Paterson view populist voters as a diverse coalition is by focusing on what they see as the economic dimension of populism. The economic factors that drive populism include: stagnant middle-class wages, rising costs of housing and other essentials, globalisation effects, and technological changes that threatening traditional jobs. Because of this focus, Russell and Paterson view someone like Bernie Sanders and his enduring popularity as an indication of how widespread populist sentiment is. As they repeatedly tell us, populism is not a “right wing” phenomenon, even though some of its most visible manifestations can be found on the political right. [At this point, it would really have helped their case if they had shifted their attention from the US to Latin America. More on this below!].

Beyond taking a closer look at the motivations of populist voters, Russell and Paterson sketch the contours of what they call “populist epistemology”. Populist epistemology combines two components. The first is scepticism about elite knowledge and elite claims to knowledge. This component has a lot in common with the critique of technocracy and technocratic knowledge in Jeffrey Friedman’s excellent book Power without Knowledge, which I have previously written about.

The second component is a defence of common sense and commonsensical knowledge. It is a bit unclear what this component really amounts to. But I think it resolves into one (or both) of two things. The first is a defence of the value of commonsensical scepticism, particularly in the context of the COVID-19 pandemic. Russell and Patterson discuss various instances where a more common sense approach might have revealed the problems with mask mandates which said you had to wear a mask in a restaurant, but could take it off to eat, or school closures that extended far beyond the point when it was clear that children were at no particular risk. The second is a defence of “local knowledge” and the validity of “lay expertise”. I’ll say more about this in a minute, but the basic idea is that local knowledge is sometimes essential to solving problems. They provide examples where expert predictions failed but lay insights proved correct, such as the case of sheep farmers who correctly predicted long-term contamination after the Chernobyl incide

Elite Reactions and Intellectual Tyranny

The final part of the book examines how elites have reacted to populist challenges. In Russell and Paterson’s view, they have reacted badly. They have doubled down on the behaviours that are responsible for the rise of populism, like elite overreach. They have also manufactured a crisis—the “post-truth” crisis—with its attendant worries about fake news and misinformation, in the attempt to explain what has gone wrong without acknowledging their part in the problem.

Against this, Russell and Paterson argue that the so-called post-truth crisis is overstated. They say that:

All the best social science research converges on the same finding: there is no evidence that we live in an age uniquely defined by misinformation or conspiratorial belief.

This is, of course, not to say that misinformation does not exist. Indeed, misinformation exists and has always existed. Their point is that it is not responsible for any of our current problems. If anything, it is the other way round: populism creates a kind of “market” for misinformation, because people who come to distrust elites go looking for “alternative experts” and alternative sources of information.

Russell and Patterson are particularly critical of what they call the "intellectual tyranny" adopted by elites, who dismiss dissent as "misinformation" rather than engaging with substantive critiques. They tell us:

The darkest side of the elite mindset is perhaps its descent into intellectual tyranny. Skepticism is mislabeled denialism; dissent is censored as misinformation or derided as conspiracy thinking; open discussion is marked off-bounds as bothsidesism.

The elite reaction to their authority being challenged has been to circle the wagons, to portray dissent as beyond the pail.

As elsewhere in the book, the COVID-19 pandemic serves as Russell and Patterson’s main example. They sketch what is more or less the standard view in the philosophy and sociology of expertise about the basis of scientific authority: scientific consensus differs from “groupthink” because it is reached via a process of contestation and debate. It is the social structures of science that confer authority on the results of the scientific process, not the brilliance and intelligence of individual scientists. But, as Russell and Patterson point out, this entails that the way in which science and the scientific consensus was portrayed during the pandemic is deeply confused. During the pandemic, there was no settled scientific consensus around masking, lockdowns, or most other public health measures that were debated. The consensus was in the process of being arrived at. You can’t treat the tentative ideas and hypotheses arrived at during the emergence of a consensus in the same way you treat a settled consensus. But this is exactly what happened throughout the pandemic.

One of the most interesting aspects of this part of the book is Russell and Patterson’s defence of “lay expertise”. These lay experts:

Often employ sophisticated data analysis and scientific methods, challenging the idea that only credentialed experts can evaluate evidence.

Russell and Patterson cite several examples of lay expertise. These include examples that many sociologists and philosophers often discuss and I have discussed in a previous piece: patient activists, Cumbrian sheep farmers, and various applications of “local knowledge”. Where Russell and Patterson differ from the existing literature is in arguing that we should view some communities of “science sceptics” in a similar way, with a particular focus on some anti-masking groups on social media. Their point here, I take it, is that, if we’re going to defend the existence of lay expertise, we can’t simply pick and choose the communities of “lay experts” who don’t challenge our pre-existing scientific and political commitments.

It must be emphasised that Russell and Patterson do not reject expertise. Rather, they advocate for epistemic humility—acknowledging uncertainty, valuing debate, and recognising that policy issues involve both facts and values. In their view, the lesson to take from the various public health missteps during the COVID-19 pandemic is not that scientists are idiots because they made so many mistakes. The lesson is rather that the fact they made some mistakes shows that some scientists were often lacking in this epistemic humility, and that it was wrong to restrict debate on things like school closures, lockdowns and masks:

Hindsight is not the correct judge of a decision, which may have rested on stable grounds when first made, but reversals made in hindsight shed an uncomfortable spotlight on why silencing dissent is so problematic.

What Russell and Paterson conclude from all this is that democracy fundamentally requires trusting citizens' capacity for political judgment rather than deferring solely to expert opinion. As they put it at the end of their book, their main aim was "to reassert a basic, moral claim about democracy: political judgment belongs to all of us". The book as a whole should be read as a broadside against the idea that we should simply defer to the elites because the elites know best.

Populism Outside the US

At this point I cannot resist a snide aside. (If you want to hear my real criticisms, skip to the next section!).

There is an extremely odd point in Russell and Patterson’s discussion of populism. At one point they say, almost in passing, that “populism across history and across the globe is a single phenomenon”. Even if this is true, they provide absolutely no reason for thinking it is true. They pay absolutely no attention to manifestations of populism across history.

Moreover, they only pay attention to manifestations of populism outside the US when they can be understood as similar to populism in the US, in the 21st century, post 2016. There is barely any discussion of populism in Latin America or Asia. There is a bit of discussion of populism in Europe, but no discussion of populism in Eastern Europe. This is quite confusing: their basic analysis of populism is that it is a reaction to globalism and neoliberalism. This is a promising basis for understanding populism across the globe (though perhaps not history). But they don't even begin to pursue it.

Instead, their story shares a common problem with the simplistic narratives around populism and misinformation that they want to criticise. It is a story that seems almost entirely based on two events in 2016 and their aftermath: the election of Donald Trump and the Brexit referendum. This is, to say the least, an impoverished basis for understanding populism—especially if you profess to think that populism “across history and across the globe” is a single phenomenon.

[A second snide aside: In their discussion of Brexit they struggle to keep the distinctions between England, (Great) Britain and the United Kingdom straight. As a professional Scottish person, few things are more guaranteed to annoy me than using “England” interchangeably with “Britain” or “the UK”].

The Problem of Terminological Slippage

Ok now for the more important problems. One of the most confusing aspects of Russell and Patterson's analysis is the terminological slippage in their use of the words "elite" and "expert." They define elites broadly, noting that:

The moniker elites, roughly speaking, refers to university-credentialed experts in high-paying, usually urban-centric jobs, alongside the politicians and members of the media who enable them.

But at various points they use the words “elite” and “expert” to refer to people who occupy a wide range of roles and positions: academics, journalists, politicians, business figures, military personnel, and simply "the rich”. It is a little difficult to see what ties all these groups together. There is no obvious connection between being a member of this broadly defined elite and being an expert in any particular domain—you don't need a university qualification to be in the military or have significant wealth.

This terminological imprecision matters because it muddies their central argument. There is a broad story they might be telling, in which populism is a backlash against the excessive power wielded by elites in general, and a narrower story, in which populism is a backlash against the excessive power wielded by an expert class within the broader class of elites. Both stories might be plausible, but they are not the same story.

The distinction becomes important when Russell and Patterson attempt to demonstrate the power of elites. They argue that:

It is implausible to hold that elites lack power because economic elites and elite groups that represent big business have enormous power to shape government policy.

This may be true, but it relies on the broader definition of "elite." What about their central target—academic elites? Do they have the same kind of power as economic elites? Some of them might because they are members of the economic elite. (I am told that professors at fancy US universities are quite highly paid). But many of them aren’t.

This conceptual looseness leads to another problem. When Russell and Patterson attribute failures during the pandemic to "elites," it is not always clear who the failure is being attributed to. For example, in their discussions of school closures and mask mandates they document numerous public policy missteps, bad political decisions, and nakedly partisan behaviours. But it is not always clear why this represents a failure with the expert class, rather than a failure with the political class, or with teaching unions, or some other group that is part of the "elite” in the broad sense but not the narrower sense.

At times it seems like Russell and Patterson’s point is not so much about the failures of the expert class as about the failure of our politicians and political systems to make proper use of the knowledge produced by the expert class. At one point they say that a core problem is that people "pick and choose the experts that tell them what they want to hear". This may well be true, but if it is true it isn’t obviously a failing on the part of the experts themselves. It’s a broader problem with our political culture.

If, as Russell and Paterson argue, expertise is weaponised, we need to ask why anyone listens to the experts—why does weaponisation work? The answer likely involves a political culture that values the veneer of expertise more than expertise itself, because it provides justification for preexisting political positions, or because it is a convenient way for a politician to avoid responsibility for their own decisions. For someone faced with the decision whether to order a nationwide lockdown, it is convenient to be able to outsource the decision to a scientist who has a model on which only a lockdown will avoid millions of deaths, rather than to take responsibility for making your own decision. If they are right, you have made the right decision; if they are wrong, you can say you were simply “following the science”.

The Neglect of Populist Politicians and Ideological Entrepeneurs

Russell and Patterson deliberately focus on populist voters rather than looking at populist politicians and the intellectuals who provide them with legitimacy. This is perfectly sensible. They are entirely right in challenging claims about populism and populist voters that are insulting and demonstrably false.

But there is an issue here.Russell and Patterson’s book spends a lot of time documenting the problems with elites and elite discourse and no time at all talking about the problems with populist politicians and the ideological entrepreneurs who court populist voters. Not only does this miss an important part of the story, it means there is a bit of a gap in their discussion of elites and the pathologies within elite discourse. It is hardly a stretch to say that populist politicians and intellectuals increasingly function like a new elite, and it is important to ask whether that new elite is any better than the old one.

Russell and Patterson spend a lot of time excoriating Tom Nichols, an American academic and writer, who has written a book called The Death of Expertise: The Campaign Against Established Knowledge and Why it Matters. I haven’t read Nichols’ book, but if Russell and Patterson’s discussion of what Nichols has to say about expertise is half-accurate, I don’t particularly want to; it doesn’t sound very insightful. But they quote Nichols as saying that populism harbours an “unfocused rage at the culture, at the elites” which produces opportunities for “savvy operators” to “exploit”.

This sounds to me like a perfect description of someone like Jordan Peterson, who is by any measure a member of the elite, and has made a lucrative career out of attacking the elite. Peterson illustrates a more general phenomenon. While Russell and Paterson may well be correct that elites weaponise expertise for political gain, they don’t discuss how other elites do the same thing in reverse. These “counter-elites” leverage anti-elite sentiment not to empower ordinary citizens, but to build their own power bases and advance their own agendas.

The lesson to take from this is that, if you want to analyse elite discourse as a form of status competition, where elites use their credentials to achieve power and status, and to advocate for their political goals, no matter how idiosyncratic or unpopular, then you need to be more even-handed than Russell and Patterson.

The Limits of Populist Epistemology

I have a few concerns with what Russell and Patterson say about populist epistemology. The concern is not that there is nothing to the idea of “lay expertise” or that they are wrong to highlight the values of “local knowledge”. The concern is more that I’m not really sure what their populist epistemology amounts to.

Jeffrey Friedman’s book, Power without Knowledge, which I mentioned earlier, serves as a useful comparison here. Like Russell and Patterson, Friedman is concerned to highlight the basic problems with technocracy. For Friedman, technocracy requires a kind of knowledge (knowledge of social problems, their causes, and how to solve them via public policy) that is incredibly hard to obtain. It is hard to obtain not because technocrats (policy makers, social scientists) are corrupt, lazy, or subject to various perverse incentives, but because social problems are complex, and evaluating their solutions is very difficult.

Friedman concludes that technocracy is doomed to fail, but he also concludes that a populist alternative is doomed to fail too. If social problems are as complex as Friedman thinks they are, common sense and commonsensical knowledge is not going to suffice to understand or solve them. Populist epistemology is no better off than technocracy. Friedman’s conclusion seems to be that we need to abandon the idea that we can leverage knowledge—whether produced by technocrats or “the people”—to solve complex social problems. [This is rather a strong conclusion and it is unclear whether he himself really endorses it. But that’s not our concern here, and in any case you don’t need to go as far as Friedman seems to in order to highlight the challenges facing technocracy].

I’m not sure whether Russell and Patterson agree with Friedman about the scale of the problem facing technocracy. In some places they seem to suggest that a populist alternative to technocracy would be preferrable. In other places they walk this back a bit. At one point they say that:

Populist epistemology—at least in its less extreme forms—is quite defensible. Very few public-policy questions rest solely on expertise: many implicate questions of values, morality, and trade-offs between different domains. The sphere where expertise can supply absolute 'answers' is really quite narrow. Most purported uses of expertise are, in fact, a bluff, where experts are trotted out to shore up the political priors of elites.

This passage encapsulates the problem. On the one hand, parts of it sound concessive: they seem to be saying that what we need is a more democratic way of doing public policy. This aligns with the mainstream view in philosophy of science, which is more concerned with the precise ways in which we need to democratise knowledge production and policy making than with arguing for the need to democratise policy making in the first place.

On the other hand, what do they mean when they say that “most purported uses of expertise are, in fact, a bluff”? Do they mean to say that there is an alternative to all this, where we get in the people who really can leverage their expertise to solve problems? If so, who are these people? Are they reformed elites, or ordinary people? How are we going to find them? Russell and Patterson simply don’t tell us.

Am I (an Elite) Just Triggered by This?

I struggled to write a conclusion to this piece. While the criticisms I have discussed may give a different impression, in fact I agree with the vast majority of Russell and Paterson's central claims. I am as sceptical of the technocratic mindset as they are (maybe more so), I am well aware of the inherent limitations of expertise, and, while some of their discussion is a little lacking in philosophical nuance and sophistication, this is more than compensated by political nuance. So why, as I read this book, did I find myself constantly wanting to disagree with them?

Maybe the reason is simply that I am one of the people they are criticising. I work at a university, I have political views, and no doubt those views have sometimes seeped into my work. I have, you might say, a vested interest in defending elites and elite institutions: my livelihood and social status, such as it is, depends on it.

Even if that's part of the reason, it isn't the whole story. To my mind, far from being too cynical about elites and the incentives that produce elite status competition, Russell and Paterson are—at least in places—not cynical enough. As I have argued, while they are happy to criticise academic experts for weaponising their expertise in the service of defending lockdowns, mask mandates, and "following the science," they don't give sufficient weight to the fact that outspoken critics of these policies were in essence doing the same thing: leveraging their expert status—or in some cases their status as outsiders, heterodox thinkers shunned by the mainstream—to agitate for their preferred approach. If we are going to view elite discourse as an elevated form of status competition, we need to be even-handed: the elites who decry the elites are as much engaged in status competition as the elites they decry.

But it goes deeper. I think the basic reason for my discontent is this. Russell and Paterson combine their negative, some might say cynical, take on the failings of elites, academic experts, and research institutions with a call for the return to older, more traditional scholarly (and journalistic) norms and ideals. We need to return, they suggest, to the old ideal of academic inquiry as open-minded and critical. Researchers should be truth-seekers and journalists should seek to report the truth. To be sure: Russell and Paterson aren't naïve. They don't claim these ideals were ever fully realised, and they don't think that researchers and academic institutions can stand "outside" of politics (as they constantly remind us, facts do not speak for themselves—they must be interpreted, and different people will interpret them differently). But they do claim that we have stopped trying to realise these lofty ideals.

What is missing from Russell and Paterson's book is a serious attempt to grapple with why we have stopped trying to realise these lofty ideals. (And why we perhaps never really tried in the first place). Why have journalists become partisan mudslingers? Why have legacy media institutions increasingly become mouthpieces for political parties and stopped even pretending to present a balance of different opinions? Why have academic researchers increasingly entered the political arena? Why do scholars favour eye-catching findings over careful, serious work? In the case of the media, I think the answer is fairly obvious: the industry is collapsing, and this is a desperate attempt, both on the part of institutions and individual journalists, to survive in a world where there is simply not enough money for the sort of careful, serious journalism that Russell and Paterson quite rightly yearn for.

In the case of universities, the answer is more complicated. It partly parallels the answer in the case of journalism: academics, particularly academics who don't study or work at universities like Harvard or Oxford, are faced with the challenge of securing, and then holding on to, a paying job, and this requires doing the things that everyone else is doing: publish (a lot), apply for funding (lots), grow your public profile (hence why I’m here). This inevitably distorts both your research priorities and the results of that research. There are of course academics who build a career by refusing to conform. But they must, of necessity, be a minority.

This can't be the full story, however, because academics who work at places like Harvard and Oxford do these things too. Indeed, they often work at Harvard or Oxford because they do these things better than everyone else. If anyone has the freedom to refuse to go along with all these perverse incentives, it is the tenured Harvard Professor.

I think the full story has something to do with the dynamics of politicisation. And here I find myself agreeing with Russell and Patterson on a crucial point: academia, scientific research, and knowledge institutions have indeed become increasingly politicised in recent decades. What was once at least nominally a pursuit of knowledge governed by scholarly norms has increasingly become a site for political battles. University departments, academic journals, and research agendas now often explicitly align with particular political positions. The scholar who claims to simply be pursuing truth is viewed with suspicion from all sides.

Given this reality, it is worth considering whether politicisation itself is the problem. It is a commonplace that politicisation is a bad thing. The commonplace is sensible enough, once we add a few caveats. Some things just are political and should remain that way. Perhaps there are some things which we don't currently view as political that really should be viewed as such. But by "politicisation" people typically mean something like "making something political that was not and should not be political." (There is a subtler variant, where something that was always in some vague sense political acquires a stronger political resonance. But we’ll ignore subtleties right now).

I don't know if politicisation is always a bad thing. If nothing else, people can disagree over what things are not and should not be political. But I do think that politicisation is sometimes a bad thing. To my mind, an under-appreciated reason why politicisation can be a bad thing is that it produces a particularly destructive dynamic. Once something (like a University or a scientific issue) becomes politicised, someone who is unhappy, or even just uneasy, about this is in a very difficult position. If they do nothing—abandon the issue or institution, or ignore its politicisation and continue as before—then, not only does it remain politicised, it retains its new political valence. Those who politicised it "win". If they try to do something about it, they cannot avoid engaging in politics, with the result that the thing in question becomes more politicised. While things can become depoliticised, this isn't that common, and it typically takes a long time. The result is that, once something becomes politicised, it tends not only to stay that way, but become more politicised over time.

We can apply this to politicised academic research. Scholarship, universities, and knowledge-producing institutions have become sites of political battles and, once that happens, it is very hard to back down. Backing down basically means letting the "other side" (or one of the many sides) win. There are of course those who want to take a more high-minded approach, equivalent to nuclear disarmament. But we know how the campaign for nuclear disarmament is doing. Those calling for disarmament in the absence of broader agreement are often seen as naïve idealists at best, or as actively undermining their own side at worst. The unilateral disarmer leaves themselves vulnerable to those who have not disarmed, just as the researcher or institution that refuses to engage in political battles may find themselves sidelined or irrelevant in increasingly partisan debates.

This dynamic helps explain why even those who recognise the problems with politicisation—like the tenured Harvard professors who could theoretically resist these trends—often don't. The costs of unilateral resistance are too high in a system where everyone else has embraced the new rules. A senior researcher who refuses to engage in the political dimensions of their field risks being dismissed as irrelevant or naïve, while watching colleagues who embrace politicisation gain influence, funding, and attention.

Russell and Paterson are right to yearn for ideals in academic research and journalism. But what they don't adequately explain is how we might come any closer to realising those ideals in a world that, for better or worse (I would certainly say worse), is heavily politicised. It's not enough to simply call for better norms when the incentives driving the current behaviour remain firmly in place. The political and economic forces that have created our current predicament—from the collapse of traditional media business models to the increasing precarity of academic employment to the growing polarisation of society as a whole—won't disappear through appeals to better behaviour alone. I must admit that, if pressed, I don’t think they will disappear—indeed, I think they will grow. I am, as a result, a good deal more pessimistic than Russell and Patterson sometimes seem to be about the future. So perhaps my fundamental issue with their book is that, far from being too pessimistic, it is not pessimistic enough. But then I would think that. After all, I am a born pessimist.

Against the backdrop of the Great Depression, mass unemployment, and general anger and hatred for politicians, Windrip, rather predictably, wins. He then establishes a fascist regime, with the help of a paramilitary force called the Minute Men who ruthlessly suppress dissent. He dismantles Congress, rips up the constitution, and creates a network of labour and prison camps.

Rather predictably, all this ends in disaster: Windrip is deposed and replaced by his campaign manager, Lee Sarason, who was the real architect of the regime. Sarason is then, in his turn, deposed by Colonel Dewey Haik, who takes the regime in a more militaristic direction, dropping the populist promises, and starting a war with Mexico. The book ends at this point, but whatever happens next isn’t going to be much fun.

I can’t say that It Can’t Happen Here is a great book. The characters are mouthpieces for ideas, their relationships don’t entirely convince, and the plot isn’t always believable. It is written in a breathless, cramped style, as if Lewis were only allowing himself to breathe at the end of each paragraph. But it is undeniably compelling. I’m not the only person who thinks so. The cover of my copy has a quote from The Guardian describing it as “eerily prescient”. This is taken from an article, written just before the 2016 US presidential election, where It Can’t Happen Here is described as “the 1935 novel that predicted the rise of Donald Trump”.

But does this make any sense? There are some superficial similarities between Trump and Windrip. They are consummate populists who play fast and loose with the facts, as well as with their campaign promises. But to say that It Can’t Happen Here has much of anything to do with Donald Trump or his government makes little sense. To describe a book written in 1935 as predicting the rise of Trump is bizarre: Trump wasn’t even born in 1935.

Moreover, It Can’t Happen Here is set in a country that bears little resemblance to the America of 2016 (or 2025): WW2 hasn’t happened yet, the Great Depression is still ongoing (this is a large part of the reason why Windrip wins), fascism and Nazism are in the ascendant worldwide (this is another part of the reason). Windrip runs on the Democrat ticket, not the Republican—this is a world where a populist Democrat is more believable than a populist Republican. Finally, if It Can’t Happen Here was a prediction, it wasn’t a very good one: America did not become a fascist regime in the 30s or 40s, and even if you think Trump shows some fascist tendencies, he hasn’t followed the Windrop blueprint terribly closely.

What I found compelling about It Can’t Happen Here is not its predictive power but the way in which it assembles many of the elements you would actually have needed for America to follow Italy and Germany in their descents into a form of totalitarianism. Once you have read it, you are in a better position to assess the claim that America is on the verge of becoming a fascist dictatorship. I don’t think this claim is particularly plausible. Reading It Can’t Happen Here helps us see why it isn’t plausible: many of the necessary elements just aren’t there.

How Does it Happen?

Lewis tells the story of Windrip’s path to power in two ways: at a distance, from the perspective of the inhabitants of a small town in Vermont, and via some brief scenes involving the key players (Windrip, Sarason, and assorted politicians and media figures in his orbit). From this we can cobble together the basic outlines.

First, the background conditions are provided by the Great Depression at home and the rise of fascism and Nazism abroad. The Great Depression provides the main background for Windrop’s rise: it created a receptive audience for his promise of $5,000 annually to every citizen, which is a centrepiece of his campaign. This promise contrasts with his main competitors, Walt Trowbridge, the fictitious Republican candidate for President who eschews empty promises that will gain him applause, and Franklin Roosevelt, who runs as an Independent after losing the Democratic primary to Windrip. But Windrip’s platform—and Lewis’s book—borrows liberally from both Mussolini’s Fascists and the Nazis, both in terms of aesthetics (the mass rallies, the military imagery) and political structures (the Minute Men are clearly modelled on Mussolini’s Blackshirts).

Second, Windrip provides the necessary charisma and personality to get people to vote for all this. He is an electrifying speaker, with the power to move an audience and mobilise the masses. More than that, though, he is an American fascist: he speaks a language that Americans can understand. He isn’t Hitler, he isn’t Mussolini, he is a homegrown product of his country, with a good line in folksy rhetoric, and an evident distrust of politics and “ideology”.

Third, the Minute Men are the enforcement arm of Windrip’s regime. These uniformed thugs wear distinctive white shirts with blue bandanas, conduct public beatings of dissenters, destroy opposition newspapers, and use systematic violence to crush resistance. They also cheat, extort, and find various ways to enrich themselves at the expense of everyone else. Like their European counterparts, they are initially voluntary but are later formalised as a state-sanctioned military force with special legal protections.

Fourth, after the election, Windrip’s regime divides the country into new administrative districts that replace the states. These districts are ruled by “Corps Commanders” who have near-absolute authority and allow the regime to bypass local government and other traditional democratic institutions. The model here, both for Windrip and Lewis, seems to be Nazi Germany and the “Gauleiter” system. The economy is reorganised into a corporate state with businesses serving the needs of government. The model here seems to be Italian fascism, where capitalists sacrificed their autonomy but held on to their property and money—so long as they were willing to serve the state.

Finally, the regime creates a network of prison and forced labour camps. They also erect a comprehensive surveillance state, with citizens encouraged to report “anti-American” activities, and completely control the media, with independent newspapers either shut down or becoming effective mouthpieces for the regime. The model here could be any totalitarian regime of the time—Soviet Russia as well as Nazi Germany.

Is This Happening Now?

In It Can’ t Happen Here it isn’t entirely clear how all of this happens, or how it happens as quickly as it does. The rise of fascism in Lewis’s America is a good deal faster than the rise of Nazism in Germany or fascism in Italy; it happens in about a week. But Lewis’s story does contain what you might view as some necessary ingredients for fascism in America:

• Mass social unrest and poverty (in the book, caused by the Great Depression and the aftermath of WW1).

• A magnetic leader around which a cult of personality can be built.

• Complete destruction of the existing legal system and democratic institutions, followed by the construction of a new system, entirely controlled by the regime.

• Reorganisation of the economy into a corporate state with business subservient to government.

• Complete state control of the press.

• Construction of a system of labour and prison camps, nominally to house political prisoners, but with no judicial oversight.

• A mass armed militia to enforce all of the above.

In America in 2025, most of these necessary ingredients are lacking. There is simply not social unrest or poverty on the scale of the Great Depression. Trump is not currently erecting a completely new system of governance, and complete state control of the press seems far off. While he is certainly mixing business with government, one often gets the impression that it is government that is subservient to business, not the other way round. Finally, there is not a mass armed militia numbering in the hundreds of thousands roaming the street to enforce compliance with the Trump regime.

In a recent article in Jacobin the historian (and host of the American Prestige podcast) Daniel Bessner argues that, while Trump’s second administration has taken an increasingly authoritarian turn, to describe it as “fascist” would be to misunderstand what is going on. America today is very different from the social and political contexts which produced Nazism and italian fascism. Interwar Europe was a period of mass social dislocation and hyperinflation. You had combat-experienced veterans roaming the streets and communism threatening to overthrow capitalism. The social conditions were ripe for fascism. None of this is the case in America in 2025.

For Bessner, Trumpism is not a foreign import. We can best understand it as an intensification of longstanding American antidemocratic traditons, rather than a rerun of Nazi Germany, or fascist Italy. Bessner highlights that Trump’s recent actions—especially the deportations of “criminals” for unspecified or fabricated crimes—build on established precedents: the long history of presidential overreach, the “theory of the unitary executive” developed under George W. Bush, and the sordid American history of arrests and deportations for crimes real and imagined. (This is not ancient history; remember what happened after 9/11). These things should trouble us, but they should trouble us because they reveal what America is and what it has been, not because they suggest that it is about to turn into Nazi Germany.

Does This Mean Everything is Ok?

No it doesn’t. Bessner’s piece is titled “This is America”. He thinks it is a mistake to draw glib comparisons between Trump’s America and Nazi Germany or Mussolini’s Italy because these comparisons provide false comfort. They foster the illusion that Trump and Trumpism is an aberration, that it is unAmerican. Bessner’s point is that it is very American—it is some of the worst tendencies from American history magnified, made explicit, brought to the surface. To refuse to acknowledge this is to refuse to understand what is happening.

In It Can’t Happen Here there is something of a tension. On the one hand, Windrip is presented as a very American fascist, as the culmination of historical tendencies that have had many different manifestations throughout American history—the Ku Klux Klan, the “Know-Nothing” movement of the 1850s, which violently opposed Catholic and Irish immigration, the Palmer Raids of 1919-1920, when the federal government arrested and deported thousands of suspected leftist radicals and anarchists. On the other, Lewis’s template for a fascist revolution borrows liberally from what was happening overseas. So which is it: homegrown American fascism or a foreign import?

Writing in the 1930s, Lewis could be forgiven for blending these influences. He was writing when fascism was ascendent globally, when America was not the all-dominant cultural and political behemoth it is today. It was perhaps plausible that, were fascism to arrive in America, it would simply be an American variant of what was happening in Germany and Italy. But Lewis clearly understood that fascism could emerge from America’s own political traditions, if the conditions were ripe, and the right person came along to exploit those conditions. Even if Trump isn’t that person, this is as true in 2025 as it was in 1935. Anyone who wants to stop this happening needs to wrestle with that fact, and this means avoiding facile comparisons with the Nazis.

My tattered library copy of Jacques Ellul’s Propaganda has a blurb from Robert R. Kirsch of The Los Angeles Times comparing the book to George Orwell's 1984. Indeed, Kirsch calls it a "far more frightening work than any of the nightmare novels of George Orwell" (we can presume he has 1984 in mind).

The comparison is fair. But it is more instructive to compare Propaganda with another work of dystopian fiction which some say inspired 1984: Yevgeny Zamyatin's We, which was first published in English in 1924. We is a more interesting book than 1984 (but I have a more general preference for Russian literature over English). Like 1984, We imagines a totalitarian state which uses all the techniques of state control to keep its citizens compliant. But, where 1984 imagines a world where citizens are brainwashed to believe things that are manifestly false (2+2=5) or plain meaningless ("ignorance is strength"), We imagines a world where citizens are brainwashed to be more rational. 1984 tells us that freedom is the ability to think that 2+2=4—that is, to think what is true. We tells us that freedom is the ability to think that 2+2=5—that is, to make up our own mind, even if our own mind is badly mistaken.

At one point in We the protagonist D-503 (a human, not a robot, but the replacement of names with numbers is of course no accident) muses:

there's nothing happier than figures that live according to the elegant and eternal laws of the multiplication table. No wavering, no wandering. Truth is one, and the true path is one. And wouldn't it be absurd if these two happily, ideally multiplied twos started thinking about some kind of freedom: that is, about some mistake?

D-503 doesn't want the freedom to make mistakes. He wants to live according to "elegant and eternal laws", which he thinks are embodied by his state. His state is good because it stops humans from being irrational—from thinking that 2+2=5.

If books can be summed up in simple lessons (they can't, if they are good, but let's ignore that) the point of We is that a totalitarian state would not be justified if it really did bring about peace, happiness and prosperity for all. It would not be justified because totalitarianism destroys human freedom. Freedom requires the ability to make mistakes, to do what is plainly irrational and contrary to one's interests. From the standpoint of freedom, there is no material difference between being brainwashed into believing the truth and being brainwashed into believing falsehoods. Being forced to do what is rational is not better than being forced to do what is irrational.

Ellul's book is about propaganda, not totalitarianism. But for Ellul there is not necessarily a difference: propaganda is a technique of social control, and the more total it is the more effective it is. Propaganda, then, is intimately connected with totalitarianism. Ellul, like Zamyatin, is worried about propaganda not because it brainwashes us into thinking things that are false (though it often does) but because it is a way of controlling us—of destroying our freedom. For this reason, Ellul is as worried about the propaganda of Western democracies (his chief examples are the US and France) as the propaganda of the Soviet Union or China. He thinks it is a mistake to think that totalitarianism in the service of democracy is any better than totalitarianism in the service of dictatorship. In either case, the fundamental problem is the same: propaganda is inimical to human freedom.

Subscribe now

Propaganda as a Sociological Phenomenon

This is Ellul’s formal definition of propaganda:

Propaganda is a set of methods employed by an organized group that wants to bring about the active or passive participation in its actions of a mass of individuals, psychologically unified through psychological manipulation and incorporated in an organization.

For Ellul, propaganda includes not just political messaging but also advertising, public relations, education, and what he calls "sociological propaganda" (sometimes “pre-propaganda”)—the general promotion of a lifestyle or worldview through cultural channels. The aim of his book is to study propaganda in all of these forms—as a sociological phenomenon.

What makes Ellul's analysis so intriguing—and disturbing—is that he approaches propaganda not as an aberration or a tool wielded by malevolent dictators, but as an inherent feature of modern technological societies. As he puts it, propaganda is "not something made by certain people for certain purposes". It is the environment in which we exist—the air we breathe, the water in which we swim.

This universality creates a puzzle, however. When Ellul discusses the most unsavoury aspects of propaganda as a technique of control, his examples are drawn from communist and totalitarian regimes (communist China, Nazi Germany, the Soviet Union). But the puzzle is only apparent. Ellul distinguishes between different categories of propaganda—pre-propaganda versus active, vertical versus horizontal, agitation versus integration. This typology helps explain why propaganda takes different forms in different societies yet, at the deeper level of analysis, is fundamentally the same phenomenon: these are all ways of gaining and maintaining social control.

Integration propaganda, for instance, which aims to get individuals to conform to society, is particularly characteristic of well-developed societies, while agitation propaganda, which aims to spark revolution or crisis, is more common in underdeveloped or febrile societies. What unites both integration and active propaganda is that they are techniques of social control, but different techniques are required for social control in different societies or in the same society at different times. For example, in Western democracies, integration propaganda operates through advertising, education, and entertainment, creating conformity not through force but through the subtle presentation of social norms. For instance, the promotion of the "American Dream" serves as a form of integration propaganda that shapes values and behaviours without appearing coercive.

The reason why Ellul so confidently proclaims the universality of propaganda is that he views it as a necessity of modern technological society. For Ellul, a technological society isn't merely one with advanced gadgets; it's a society organised around technique, efficiency, and instrumental rationality. Any technological society has advanced physical infrastructure—roads, hospitals, schools, airports, etc. Propaganda is the counterpart to this physical infrastructure—it is a means of ensuring compliance from the members of a modern technological society. Compliance requires more than just simply “going along” with things. It requires psychological integration. For Ellul, the totalising nature of propaganda in authoritarian regimes simply represents the most explicit manifestation of tendencies inherent in all modern societies, regardless of their economic or political organisation.

This is what makes Ellul's analysis more frightening than Orwell's. In 1984, there is a clear villain (Big Brother, the State). In Ellul's analysis, propaganda has no single source—it is woven into the very fabric of technological society:

Propaganda is called upon to solve problems created by technology, to play on maladjustments, and to integrate the individual into a technological world... In the midst of increasing mechanization and technological organization, propaganda is simply the means used to prevent these things from being felt as too oppressive and to persuade men to submit with good grace.

Propaganda is simply the form that social control takes in an advanced society.

The Necessity for Propaganda in Modern Society

This claim—that modern democratic states require propaganda to function—is one of Ellul’s most provocative claims. Why does he think this? This necessity arises from several factors.

First, there's the problem of governing in a mass society where public opinion matters:

Propaganda is needed in the exercise of power for the simple reason that the masses have come to participate in political affairs.

Democracies face a dilemma: governments need citizen support, but citizens often lack the knowledge, time, or inclination to form well-reasoned opinions on complex policy matters. What's the solution? For Ellul, there is only one possible solution:

The democratic State, precisely because it believes in the expression of public opinion and does not gag it, must channel and shape that opinion.

We can see this clearly in the way that democratic states manage public opinions on complex scientific issues or in economic policy. Citizens are not simply presented with technical data and left to make up their own minds. Instead, they create narratives that simplify issues and attempt to direct public action towards predetermined goals (recycling more, getting vaccinated). For instance, public health campaigns during the Covid-19 pandemic didn't just provide information—they attempted to direct behaviour by using emotionally laden messaging like “protect the vulnerable”. This approach is clearly not just about informing; it is about channelling public opinion toward specific policy outcomes.

Second, there's an arms race dynamic. When enemy states (or competing corporations) use propaganda, the most effective defence is counter-propaganda. This creates a spiral where propaganda becomes ubiquitous simply because everyone else is using it. As Ellul observes:

When enemy countries use propaganda, the most effective weapon to fight back with is more propaganda. This leads to a sort of arms race, where everyone has reasons to use propaganda because everyone else is using it.

This pattern continues today in the information competition between Western democracies and countries like Russia and China.

Third, and perhaps most disturbingly, Ellul suggests that modern humans psychologically need propaganda:

it [propaganda] succeeds primarily because it corresponds exactly to a need of the masses... the need for explanations and the need for values... Effective propaganda needs to give man an all-embracing view of the world, a view rather than a doctrine.

In a complex world where traditional meaning-making structures have weakened, propaganda offers comforting simplifications:

The great force of propaganda lies in giving modern man all-embracing, simple explanations and massive, doctrinal causes, without which he could not live with the news. Man is doubly reassured by propaganda: first, because it tells him the reasons behind the developments which unfold, and second, because it promises a solution for all the problems that arise, which would otherwise seem insoluble.

Consider how conspiracy theories function in contemporary society—from QAnon to anti-vaxxers. These aren’t simply systems of belief. They provide adherents with comprehensive worldviews that explain complex problems through simple narratives of good versus evil. Conspiracy theories offer a sense of meaning and purpose to those who feel adrift in a complex, chaotic world. Something similar could be said about certain political ideologies, like Marxism, which provide simple, all-encompassing explanations for societal problems.

One interesting feature of this diagnosis is that the problem it adverts to, and which propaganda is supposed to solve, sounds a lot like what is nowadays called the “meaning crisis”:

The individual who feels himself in conflict with the group, whose personal values are different from those of his milieu, who feels tension toward his society, and even toward the group in which he participates—that individual is in a tragic situation in modern society. Until recently, such an individual enjoyed a certain freedom, a certain independence, which allowed him to release his tension in external—and quite acceptable—actions. He had a circle of personal activities through which he could express his own values and live out his conflicts … But in the technological society, the individual no longer has either the independence or the choice of activities sufficient to release his tensions properly. He is forced to keep them inside himself.

This sentiment—that atomised modern society has reduced our freedom, that the individual is lost in the mass—is common nowadays, particularly among conservatives who long for a more religious past (Ellul clearly has some of these sympathies, but he’s the sort of interesting thinker who is hard to characterise politically).

I buy the idea that there’s a genuine social and psychological problem here. But I really do not buy the idea that things were somehow better in the past. One of the dominant themes in fiction written in and about the past is precisely how confining older, more traditional forms of society were, particularly for people without independent means.

To pick some examples from a literature with which I am very familiar, it is a common theme in Scottish literature that deals with rigid class hierarchies (think of William McIlvanney’s Docherty, set in a mining community in the early 20th century), or with the challenges that an independent-minded woman in rural Scotland faced if she wanted to enjoy her independence (Nan Shepherd’s The Quary Wood, or Lewis Crassic Gibbon’s Sunset Song). The characters in these books strive for an independence they cannot achieve precisely because of the rigidity of traditional society.

There is an interesting point here, but it’s a bit different to the one Ellul makes. It’s not so much about the loss of a certain kind of freedom, which was never widely enjoyed in the first place (maybe young men of independent means were genuinely more independent in the past; who knows?). It’s more about a shift in the system of control. Putting things very broadly, in traditional society people are trapped by their assigned roles in a rigid hierarchy, but the bonds are more immediate—to family, church, the land, and so on. This is why someone like Chris Guthrie in Sunset Song or Martha Ironside in The Quarry Wood experiences a deep tension: pursuing independence means abandoning home and family. In Ellul’s modern technological society, we’re no less trapped than we were before, but everything is at a far greater remove, and we can no longer even begin to identify with that which binds us.

Truth, Rationality, and Propaganda

Another important aspect of Ellul's analysis is his attack on the idea that propaganda is necessarily based on falsehoods. Contrary to popular belief, he argues, effective propaganda often uses facts and appeals to rationality:

there is such a thing as rational propaganda, just as there is rational advertising. Advertisements for automobiles or electrical appliances are generally based on technical descriptions or proven performance—rational elements used for advertising purposes. Similarly, there is a propaganda based exclusively on facts, statistics, economic ideas. Soviet propaganda, especially since 1950, has been based on the undeniable scientific progress and economic development of the Soviet Union; but it is still propaganda, for it uses these facts to demonstrate, rationally the superiority of its system and to demand everybody's support.

In fact, as society progresses, propaganda becomes increasingly fact-based:

the more progress we make, the more propaganda becomes rational and the more it is based on serious arguments, on dissemination of knowledge, on factual information, figures, and statistics.

It is therefore simply a mistake to view propaganda as necessarily connected with falsehood, lies, and deception. Propaganda in the form of blatant falsehoods is simply a symptom of insufficiently advanced society. The more developed a society, the more likely it is to utilise rational propaganda.

Here we find a connection with Walter Lippmann's analysis of public opinion, which prolific Substacker Daniel Williams has written about at length. In Lippmann's analysis, citizens face the fundamental problem of forming accurate "pseudo-environments" (mental representations of reality) in a world that is "too big, too complex, and too fleeting for direct acquaintance." They must rely on simplifying stereotypes to make sense of the world, and these inevitably distort the world, because a simple picture of the world is never going to be terribly accurate.

Where Lippmann diagnoses the epistemological challenges of forming accurate beliefs about politics, Ellul reveals the institutional mechanisms that exploit these limitations. If Lippmann shows why our pictures of the world are inevitably simplified and distorted, Ellul demonstrates how propaganda systematically shapes those simplifications to serve institutional ends. Lippmann helps us see that we need interpretive frameworks to make sense of complex realities; Ellul shows us that whoever controls those frameworks wields immense power.

This suggests something deeply troubling: that in complex modern societies, the formation of public opinion is never truly free from manipulation. The "pseudo-environments" that guide our political behaviour are not just shaped by cognitive limitations but actively engineered. This is why Ellul sees propaganda as so fundamentally at odds with freedom—it exploits the very cognitive processes that make thinking possible in the first place.

The Psychological Effects of Propaganda

Ellul identifies several psychological effects of propaganda. The first is what he calls "psychological crystallization":

Propaganda furnishes objectives, organizes the traits of an individual's personality into a system, and freezes them into a mold... prejudices that exist about any event become greatly reinforced and hardened by propaganda; the individual is told that he is right in harboring them; he discovers reasons and justifications for a prejudice when it is clearly shared by many and proclaimed openly.

The result is cognitive efficiency that comes at the cost of nuance:

he now has a supply of ready-made judgments where he had only some vague notions before the propaganda set in; and those judgments permit him to face any situation.

Perhaps most concerning, propaganda creates a form of alienation:

In large groups, man feels united with others, and he therefore tries to free himself of himself by blending with a large group. Indeed, propaganda offers him that possibility in an exceptionally easy and satisfying fashion. But it pushes the individual into the mass until he disappears entirely.

This pushing of the individual “into the mass” is crucial to the success of propaganda. Propaganda often creates what Ellul calls “politicisation” and what we would call polarisation:

A modern State can function only if the citizens give it their support, and that support can be obtained only if privatization is erased, if propaganda succeeds in politicizing all questions, in arousing individual passions for political problems, in convincing men that activity in politics is their duty.

But propaganda can simultaneously function as an "agent of privatisation" when opposing propagandas create cynicism and scepticism. Indeed, some propaganda deliberately aims at privatisation, emphasising the complexity of political questions and the need to leave them to "experts," encouraging citizens to disengage from politics entirely. Which tactic is employed depends entirely on the propagandist's goals in a particular situation.

This analysis of how propaganda leads to politicisation might look like it anticipates contemporary concerns with social media echo chambers, the algorithmic reinforcement of beliefs, and polarisation. And, in a way, it does. But the fact that it does shows why these problems which are supposedly due to social media are not and cannot really be due to social media. Social media is simply a new(ish) technology which provides a mechanism by which we can surround ourselves with confirming voices. But the basic psychological tendencies on which this mechanism operates haven’t changed. We have a general human need to surround ourselves with confirming voices.

Propaganda and Democracy

The most important aspect of Ellul’s analysis is what he has to say about propaganda and democracy. For Ellul, democracies need propaganda to function, but propaganda undermines the foundations of democracy.

Ellul is not satisfied with the common claim that propaganda is harmless when used to support democratic (or other “good”) values. He rejects as "silly" the idea that propaganda is good simply because it serves democracy:

Once democracy becomes the object of propaganda, it also becomes as totalitarian, authoritarian, and exclusive as dictatorship.

More fundamentally:

democracy is not just a certain form of political organization or simply an ideology—it is, first of all, a certain view of life and a form of behavior... But if democracy is a way of life, composed of tolerance, respect, degree, choice, diversity, and so on, all propaganda that acts on behavior and feelings and transforms them in depth turns man into someone who can no longer support democracy because he no longer follows democratic behavior.

In other words, the means (propaganda) corrupt the ends (democracy). The state that uses propaganda to uphold democracy inevitably undermines the very democratic values it claims to defend:

the democratic State, even if it does not want to, becomes a propagandist State because of the need to dispense information... It is, in effect, a State that must proclaim an official, general, and explicit truth. The State can no longer be objective or liberal... It can no longer tolerate competition, because a State that assumes this function no longer has the right to err; if it did, it would become the laughing stock of the citizenry, and its information would lose its effect, together with its propaganda. For the information it dispenses is believed only to the extent that its propaganda is believed.

During the Covid-19 pandemic, governments felt compelled to present unified messaging about protective measures, even when scientific understanding was evolving (or just plain wrong). The need to maintain credibility often led to oversimplifications or delayed admissions of changing evidence. This demonstrates Ellul's point that even democratic states become "propagandist States" when dispensing information becomes central to governance.

This puts modern democratic states in a paradox from which they cannot escape: they must use propaganda to function and survive, but in doing so, they become less democratic. We see this tension play out in contemporary debates about regulating social media or combating "disinformation." Democratic states feel compelled to establish what counts as truth, yet in doing so, they risk becoming precisely the kind of propagandist state that Ellul warned against.

Propaganda and Freedom

We can now see why you might think that Ellul's analysis is more frightening than Orwell's dystopian vision. In 1984, power is naked and brutal. But there is at least the possibility of resistance because the source of oppression is clear. In Ellul's analysis, propaganda operates not just through crude lies but also through facts, reason, and even our desire for meaning. It works not just through state media but through education, advertising, and culture. Most unsettlingly, it succeeds because it offers something we want—certainty, simplicity, belonging.

Ellul's final warning remains powerful:

The question is not to reject propaganda in the name of freedom of public opinion—which, as we well know, is never virginal—or in the name of freedom of individual opinion, which is formed of everything and nothing—but to reject it in the name of a very profound reality: the possibility of choice and differentiation, which is the fundamental characteristic of the individual in the democratic society.

Like Zamyatin's We, Ellul reminds us that freedom must include the freedom to be wrong—to reject the elegant and eternal laws, to wander and to waver. It's the freedom to think that perhaps 2+2=5, even if we know better. Propaganda, by destroying this possibility of choice, undermines not just democracy but humanity itself:

It creates a man who is suited to a totalitarian society, who is not at ease except when integrated into the mass, who rejects critical judgments, choices, and differentiations because he clings to clear certainties.

What can a book written in the 60s tell us about contemporary politics? It cannot give us any simple or clear lessons: it is a mistake to just transplant a book, even an insightful one, that was written in one context to another, very different, context. But, if you are at all sympathetic to Ellul’s analysis, you should be concerned about some features of contemporary politics and political discourse.

It doesn’t take an analysis to explain why we should be worried about lies, misinformation, and disinformation. The value of an analysis like Ellul’s is that it explains why we should also be worried with the way in which mis and disinformation have been labelled social problems—indeed, some say, the most pressing social problems of the day.

One reason we should be worried is that, as Ellul has shown us, it is simply a mistake to think that propaganda deals in falsehoods. Effective propaganda typically uses facts, but it uses facts in a way designed to produce action and conformity. The skillful propagandist doesn’t distort the facts; they selectively highlight the facts in a way designed to produce the behaviour that they want to produce. One politician might highlight rising rates of specific crimes in certain areas to justify stricter policing and sentencing. Another might emphasise the overall downward trend in violent crime over decades to argue for criminal justice reform. Neither is lying—they're selecting different facts from the same dataset to mobilise different constituencies. As Ellul observes:

The propagandist doesn't lie when they can be caught in the act because that would be disastrous to their aims; they only lie when there is no chance of being contradicted.

Another, perhaps deeper, reason is that many of the strategies that have been proposed for counterring misinformation—for example, pre-bunking or “inoculation”—look a lot like attempts to combat propaganda through counter-propaganda. As Ellul might observe, when democratic institutions set themselves up as the arbiters of truth, and set about priming citizens to think in the ways those institutions would like them to think, they illustrate the basic tension between propaganda and democracy. A democratic state—or, at least, a liberal democratic state—cannot set itself up as an arbiter of truth without undermining the fundamental principles on which it is based.

The problem, as Ellul also points out, is that the alternative is to let others—other states, or those within democratic states who are hostile to them—set themselves up as arbiters of truth. Thus Ellul tends towards a kind of cynicism: democracies simply cannot win. But he prefers cynicism to refusing to face reality:

propaganda, regardless of origin, destroys man's personality and freedom... the illusion that one engages in psychological action as a defense, while respecting the values of democracy and human personality, is more pernicious than any cynicism which looks frankly at the true situation.

Even if you think Ellul’s analysis of propaganda goes too far and tends towards the dramatic, it is hard to read his book without agreeing that there is something here, some truth to which he is pointing. We should face this truth rather than pretending it simply doesn’t exist.

“This book defends a conclusion that I sincerely wish were false … we have very little political knowledge” (p. 10).

What does “political” mean here? Roeber’s definition is, read literally, quite strange:

“By ‘political,’ I mean something specific. I mean useful to either the Democrats or Republicans for capturing or maintaining political power” (p. 10).

If we assume, charitably, that Roeber recognises that people who don’t think about American politics have political beliefs, then we get a more sensible definition: political knowledge would be knowledge that is useful to political parties or factions for capturing or maintaining political power. His claim is that we have very little of this sort of knowledge.

This is a conclusion with which I, by and large, agree; indeed, I have defended an argument for scepticism about political knowledge which is very similar to Roeber’s. But I have come to be a bit dissatisfied with the route by which I—and by extension Roeber—reach our sceptical conclusions. Let me start by explaining how Roeber reaches his sceptical conclusions, before explaining why I don’t think his argument can take him as far as he thinks it can. (If you want a slightly different take on Roeber’s book, check out this post by my fellow Substacker Michael Hannon).

Subscribe now

Roeber’s Argument

Roeber’s argument is very simple.

First, we form our political beliefs by relying on testimony (on what others tell us about political matters). The “expert” (the economist, the political scientist, the statistician) is not in a fundamentally different position from the ordinary citizen. Any expert is going to have to rely on the testimony of other experts. If I want to get statistics on immigration and its impact on the economy, I need to rely on data, reports, and other forms of evidence produced by others, and the people producing that data will, in turn, need to rely on others. (And so on). The same goes for the journalist, for whom “fact-checking” really just means checking one report of what “the facts” are against other reports.

Second, political testimony is unreliable. There are several reasons why political testimony is unreliable. The first is that people often have an incentive to lie or bullshit about politics: to say things we know are not true (lying) or don’t care whether they are true (bullshitting) in order to advance our political aims as we see them (to defend our favoured party, to destroy our political enemies in debate, to convince people to vote the way we want them to). The second is that it is hard to have true political beliefs even if we sincerely care about the truth of our political beliefs: truly understanding things like the economy, immigration or foreign policy is difficult. The third is that, as Roeber puts it, we are “incapable of thinking straight about political issues” (p. 52): we are subject to myriad biases and prone to identity-protective cognition or motivated reasoning (reasoning in ways designed to preserve our political identities). Both the second and the third reasons lead to the same conclusion: political testimony is unreliable because many of our political beliefs are simply false.

Third, we lack the ability to discriminate between reliable and unreliable political testimony. While Roeber spends less time on this part of his argument than the first two parts, it is important. The fact that political testimony is unreliable wouldn’t matter that much if we were good at discriminating between reliable and unreliable political testimony. But Roeber thinks we are not good at it, for much the same reasons that he thinks that political testimony is unreliable: we aren’t very good at detecting lies and bullshit, discriminating between reliable and unreliable political testimony is also hard, and—perhaps most importantly—we are often motivated not to really try. Rather than setting aside our political convictions and objectively evaluating political testimony, we tend to pick and choose according to whether the testimony supports our existing political convictions (which, remember, Roeber thinks are likely to be wrong).

Roeber concludes that we have very little political knowledge. As he puts it:

“People simply aren’t trustworthy when they are talking about politics. They aren’t trustworthy as speakers, because they’re too likely to say false things, and they aren’t trustworthy as listeners, because they are too likely to believe false things” (p. 61).

You (yes you, the reader) might think “quite right; people are really stupid when it comes to politics”. You might think this because you think you are less prone to bias and identity-protective cognition than other people. Or you might think this because—at least when it comes to certain political matters—you are an expert. But Roeber intends his conclusion to apply to everyone. Or at least he intends it to apply to almost everyone. He certainly thinks it applies to anyone who is a partisan of a political party, and that includes many subject-matter experts.

I think that Roeber’s conclusion is a bit too strong. I’m not convinced that his argument can be extended to experts as easily as he thinks it can, and I suspect there are quite a lot of people who do not approach politics in the way he seems to assume (political partisans are ubiquitous in political discussions and debates because political discussions and debates attract political partisans, not because everyone is a political partisan). Still, I think his argument might support a weaker conclusion that is still interesting and worth considering: we have far less political knowledge than we typically assume. In the rest of this post I will raise two concerns for Roeber’s argument, which will arise whether we use it to establish my weaker conclusion or his stronger one.

Two Kinds of Political Scepticism

While Roeber has several reasons for thinking that we have very little political knowledge, it is very clear that the main reason is that we do not inhabit a political climate that is conducive to “thinking straight” about politics. At least in the contemporary US (and Roeber’s book is incredibly focused on the contemporary US), political issues are closely connected with questions of cultural and social identity, with the result that we typically engage in politics as a way of preserving and maintaining our cultural and social identities.

For what it’s worth, I think he’s largely right about this, though at least outside the US it doesn’t necessarily take the form of aligning with specific political parties (I know a lot of people with leftish politics—I’m an academic—but I don’t know many people who really *identify* with the UK Labour party, or for that matter with any political party). Anyway, the upshot of all this is that Roeber’s defence of scepticism has a hopeful note to it:

“our political speech would be much more reliable if we did not identify with our politics. In a political climate where we do not identify with our politics, trusting someone who is talking about politics might resemble trusting someone who is telling you her name more than it resembles trusting someone in a poker game. In this political climate, there would be much less reason to worry about our political beliefs” (p. 77).

If the main reason why we lack political knowledge is simply that we identify too much with politics, then it stands to reason that, in a world where we identify with politics less, we will have more political knowledge.

I must admit I am not really sure what it would mean to identify with politics less. If political knowledge is fundamentally knowledge that is relevant to holding on to or gaining political power, then refraining from identifying with politics would be tantamount to refraining from caring about whether you have any political power; not an attractive option for many people, I suspect. But perhaps Roeber has something a little easier in mind. The book very much gives the impression that the problem is not so much caring about politics as identifying with political parties, so perhaps it’s enough to tear up your membership card and stop viewing being “into politics” as akin to supporting a football team.

Anyway, my real issue with this is that I’m not convinced that the main reason we lack political knowledge is that we identify too much with politics. I am more inclined to think that the main reason is simply that political knowledge—or at least the sort of political knowledge that Roeber focuses on in his book—is just hard to obtain.

Roeber’s view seems to be that the fundamental difficulty in political epistemology is not with the subject matter (the body of facts out there waiting to be understood) but with the subject (the person or people trying to understand the facts). My view is that, while there are difficulties on the subjective side, the fundamental difficulty is with the subject matter. Political knowledge is (or would be, if we had any) knowledge of the social world—knowledge of human relations, of how societies work, of how the economy works, and so on. I think that knowledge of the social world is orders of magnitude more difficult to obtain than knowledge of the natural world.

One of Roeber’s recurrent examples of (the failure of) political knowledge is knowledge about the impact of increasing the minimum wage on employment rates. The way Roeber tells it, the main obstacle to knowledge here is that one side will interpret the evidence (whatever the evidence is) in such a way as to support increases to the minimum wage whereas the other side will interpret the evidence in such a way as to not support an increase. But another obstacle is the fact that most people haven’t got a clue what the evidence is, much less evaluate it. For the economists, who presumably can evaluate the evidence, there is the difficulty of fitting economic models in which, when the minimum wage goes up, employment goes down to the real world where real actors (employers) might not react to increases in the minimum wage in the way that economists expect. Even if someone can work through all these difficulties and achieve knowledge, it is not going to be easy, even setting aside any bias on the part of economists. (This is a simplified version of an argument from this excellent book).

Another of his recurrent examples is knowledge about the impact of immigration on the economy. Again, the way Roeber tells it, the main obstacle is that each side will interpret the evidence in a way that fits with what they already think. One side will argue, citing various bits of evidence, that immigration is bad for the economy, or at least for certain parts of the economy; the other will argue, citing different bits of evidence, that it is good for the economy, or at least certain parts of it. (The fact that these views are not necessarily inconsistent with each other will not occur to anyone on either side). But, again, another obstacle is the fact that most people haven’t got a clue what the evidence is, or how to evaluate it. Those who can evaluate the evidence are going to face similar issues to economists when they try to compare their economic models with the real world. Even if they can figure it all out, it will take a lot of work to do so, even setting aside any bias on the part of the experts.

A good way to distinguish between Roeber’s view and mine would be to imagine what would happen if we get to the point Roeber would like us to reach: our political climate is now one in which people no longer identify with their politics, or at least they do not do so to anything like the same extent. Is this a world where we have a lot more political knowledge than we do right now? I would certainly expect there to be a lot more agreement on certain political issues. But your views on whether there will really be a lot more political knowledge are probably going to align with your views about the prospects for the social sciences becoming more like the natural sciences. As a sceptic, albeit a somewhat half-hearted one (hence why I prefer my weaker conclusion to Roeber’s stronger one), my best guess is that you’d have more political knowledge in this imagined future, but not a huge amount more. This is because the fundamental obstacle to political knowledge is not our propensity to biased and motivated reasoning but the inherent difficulty of knowing anything about the social world.

Political Humility

Roeber thinks there are important practical consequences that follow from his conclusion. First, because we have very little political knowledge, we should have very few political beliefs. The basic thought is intuitive: if, after reading Roeber’s book, you are convinced by his argument that it is very difficult to get political knowledge, you should abandon your existing political beliefs (because they are very likely false) and refrain from forming any new ones (because they will most likely be false).

Second, this does not mean we should refrain from engaging in politics. But it does mean that we should engage in politics in a different way. We should engage carefully, and with a high degree of humility. For example, we shouldn’t go around loudly proclaiming our political beliefs (because we shouldn’t have many political beliefs). But we can go around making what Roeber calls “educated guesses”: instead of saying “these tarifs will seriously damage the economy” we should say things like “I think that these tarifs will seriously damage the economy, based on what I’ve read by various economists”. We also might want to consider whether we should vote in elections, if none of the candidates have platforms that appeal to us, and focus on forms of political action other than voting (or telling people what we think).

It is hard to disagree with a lot of this. We are prone to over-confidence, especially when it comes to politics. It is hard to spend much time on social media, or at least the corners of social media where people talk about politics, without recognising that one of the basic problems is that the people who spend the most time talking about politics often know very little about anything. Or at least very little about substantive political issues. A lot of “talking about politics” is really just talking about what other people are saying about politics: the stupid thing your political opponent just said, the media circus around certain prominent politicians, the tweet that expresses an opinion so stupid or downright mad that no reasonable person could really think it.

One objection to Roeber’s call for more humility is that it will be heeded by those who don’t need to hear it, and ignored by those who do. As W.B. Yeats wrote, in more poetic language:

“The best lack all conviction, while the worst / Are full of passionate intensity."

The post on Roeber’s book I mentioned earlier, by Michael Hannon, expands on this point. I agree with Michael that we might need the best to retain a bit of arrogance, if only to balance out the passionate intensity of the worst.

Another objection, and the one I want to focus on, concerns how we think about humility. Roeber seems to think of humility as a matter of the degree of confidence with which we hold our political attitudes: the humble person may have some political opinions—or educated guesses—but they don’t have much confidence in them. If they’re a Bayesian, they might attach figures to their attitudes and say things like “I’m 0.6 on lab leak”. If they’re normal, they’ll just say things like “I think it is likely that Covid escaped from a lab, but I’m not sure”. What they will not say is something like “We know that Covid escaped from a lab” or just “Covid escaped from a lab”.

This is a perfectly sensible way to think of humility, but it doesn’t lend itself well to practical advice. The problem is that, if trying to be humbler means trying to lower your degree of confidence, it is very hard to properly calibrate your degrees of confidence. You need to avoid having a degree of confidence that is too high, given your actual evidence, but you also need to avoid having a degree of confidence that is too low, given that evidence. Roeber’s book is addressed to someone who is in the first position: they are far more confident that, say, immigration is good for the economy than they should be given their actual evidence. But I worry that someone who is convinced by his arguments is liable to become less confident of all sorts of things than they should be.

Imagine someone who is encountering these sceptical arguments for the first time: they have never even considered that their thinking about politics might be biased, or driven by their social and cultural identity. If they are convinced by Roeber’s arguments, they are liable to go too far in the other direction—to throw the baby out with the bathwater, so to speak, and give up on the prospects of having any good grounds for their political attitudes.

Let me suggest an analogy to try and make this point clearer. (It is quite a close analogy, so hopefully this will work). Think back to when you first heard about the heuristics and biases tradition in psychology: confirmation bias, the availability and representativeness heuristic, and all that stuff. If you were anything like me, you jumped from accepting (if only implicitly) a naively optimistic picture of human rationality to accepting an overly pessimistic picture. It takes time—and a good bit of careful thinking—to arrive at a more measured view of things. Even then, it is hard to be particularly confident that you have got the balance right. Calibration is hard, both when it comes to taking on board the lessons of the heuristics and biases tradition and when it comes to recognising the human propensity to engage in identity protective cognition.

There is another way of thinking of humility that might avoid this problem. We can think of humility not in terms of having a low degree of confidence in your beliefs (or guesses) but rather in terms of the stance you take towards those beliefs. On this understanding, the humble person can have many political beliefs, held with whatever degree of confidence they think is merited, but recognise that many of those beliefs may well be false. Crucially, though, this is not because they can pinpoint which of these beliefs are false (if they could, they would be in the somewhat incoherent position of believing both that, say, immigration is good for the economy and that their belief that immigration is good for the economy is probably false). They know that many of their beliefs are likely false, but they can’t say which ones.

This sort of humility may well be better supported by Roeber’s argument than the sort he advocates for. Roeber’s sceptical arguments are based on general considerations about political beliefs: how we get them, why they will often be false, and so on. They are not, on the whole, based on reasons for thinking that any particular political beliefs are likely to be false. This makes it difficult to know what we should conclude about the particular political attitudes we happen to hold. Should I, after reading Roeber’s book, downgrade my confidence in all my political attitudes by the same amount? Should I look at each attitude individually and decide what to do with it? Far simpler to instead just recognise that, while I may have my political beliefs, many of them are very likely going to turn out false, but not for reasons that really help me to adjust my degree of confidence in any particular attitude.