Can democracy work?

In "Public Opinion" (1922), Walter Lippmann argued that the vastness, complexity, and invisibility of the modern world make democracy impossible. He got a lot right.

“It was no part of political science… to think about how knowledge of the world could be brought to the ruler. . . . What counted was a good heart, a reasoning mind, a balanced judgment. These would ripen with age, but it was not necessary to consider how to inform the heart and feed reason. Men took in their facts as they took in their breath.”

—Walter Lippmann, Public Opinion

I. Knowledge, Decisions, Democracy

Good decisions depend on knowledge. If you cross the road and don’t know a car is hurtling towards you, you’ll be killed. If you invest money on the basis of wishful thinking, you’ll become destitute. If you marry an awful person you think is kind and trustworthy, you’ll be miserable.

It's plausible that this generalises to politics. If societies are governed by ignorant or misinformed rulers, they’ll produce bad laws and implement bad policies. For example, Mao believed that plants behave like uncompetitive communists, which led him to demand stupid farming practices. These practices played a role in famines that killed tens of millions of people.

In democracies, there’s an important sense in which voters are ultimately the rulers. Given this, it’s plausible to think that democracies will only make good decisions—pass good laws, implement good policies, and so on—if citizens are well-informed.

On the surface, this looks challenging. How could the average voter in a modern democracy become sufficiently well-informed about topics as diverse and complex as climate change, crime, economics, trade, international relations, and so on?

Perhaps this is the wrong way to think about it. In most democracies, it’s rare for citizens to vote on specific policies. Instead, they outsource their decisions to representatives. Given this, perhaps citizens don’t need to be well-informed about politics. They just need to be well-informed about which representatives are likely to make good decisions.

This doesn’t really help, though. If citizens aren’t knowledgeable enough to make good decisions, how could they be knowledgeable enough to decide which representatives would make good decisions? Further, these representatives aren’t a special breed of human beings. They’re just other citizens, confronted with the same seemingly impossible task.

II. Epistemic Critiques of Democracy

Democracy sounds good. It sounds so good that dictators characterise their societies as democracies and run sham elections. North Korea is a totalitarian state run by psychopathic thugs who murder, starve, and imprison the population. What’s it called? The Democratic People’s Republic of Korea.

But is democracy good? Although many people treat democracy as sacred today, historically many intellectuals and philosophers thought the answer to this question is “obviously not”.

There are many critiques of democracy, but most of them are fundamentally epistemic. That is, they argue that democracy is bad because voters lack the knowledge, wisdom, or understanding required to make good decisions.

Some of these critiques identify contingent problems. For example, left-wing intellectuals often argue that voters are only misinformed as a result of elite manipulation. On this view, propaganda, “ideology”, and other sinister forces drive the hoi polloi to adopt beliefs at odds with their own interests. If you could get rid of those things, voters would see reality objectively (i.e., as left-wing intellectuals see it) and everything would be fine.

Others take a less favourable view of the general public. At least since Plato argued that ordinary citizens inevitably lack wisdom and are too easily led astray by emotions, a prominent critique of democracies focuses on the ignorance and irrationality of voters.

This critique seems to be vindicated by a vast body of research showing that most voters are shockingly uninformed about even extremely basic political facts. Moreover, although some citizens are more knowledgeable and politically engaged—if you’re reading this, that’s almost certainly you—they tend to process information in extremely biased, partisan ways. (That’s probably you, too).

In other words, most voters are ignorant, but some are just irrational.

Is this inevitable? Some intellectuals are optimistic. They argue that if we had better education, or more democratic participation, or more thoughtful and open-minded forms of political dialogue, then we could cure the citizenry of ignorance and bias.

Others are more pessimistic. They argue that the basic problem is one of incentives. In large democracies, individual votes have basically no impact on policies. Given this, voters have no incentive to become well-informed or to reason in a reasonable way. In rationally responding to incentives, they therefore opt for ignorance or unreasonableness instead.

However, this critique rests on an optimistic assumption: namely, that if citizens were rational, and if they took the time to become well-informed, then democracy could work.

Is that true?

III. Walter Lippmann’s Radical Critique of Democracy

In my view, Walter Lippman’s (1922) Public Opinion is the most important work of political epistemology (i.e. the study of how politics interacts with questions about knowledge and ignorance) ever written. Along with his follow-up The Phantom Public (1925), it’s also the most radical and interesting epistemic critique of democracy.

According to Lippmann, it’s a mistake to think that the fundamental epistemic problem with democracy is that citizens are uninformed, misinformed, irrational, or the victims of propaganda. Of course, he doesn’t deny that such things exist and are bad. However, he argues that citizens in modern democracies confront insurmountable obstacles to understanding society and making good decisions even if they try to become well-informed in the absence of all such distorting factors.

The modern world, argues Lippmann, is so vast, complex, and inaccessible that voters—even rational, well-meaning voters who consume well-meaning media—cannot understand it.

More specifically, Lippmann draws a distinction between two things:

1. The real environment: Those features of reality that are relevant to political decision making in modern democracies. This includes the economy, social relations, human behaviour, international conflicts, technological developments, the environment, and much, much more.

2. The pseudo-environment: The mental model or “picture” that citizens form in an attempt to understand and navigate the real environment.

Lippmann’s central thesis is that the pseudo-environment is inevitably a profoundly selective, simplistic, and distorted representation of the real environment. Given that we make decisions on the basis of such pseudo-environments, there is no reason to expect good decisions.

“The way in which the world is imagined,” he writes, “determines at any particular moment what men will do. It does not determine what they will achieve. It determines their effort, their feelings, their hopes, not their accomplishments and results.”

Why does Lippmann think the world we imagine—the pseudo-environment—must be so radically different from the real environment?

IV. The real environment

Writing at the beginning of the twentieth century, Lippmann argues that his contemporaries do not appreciate the sheer vastness and complexity of the modern world. This is a world composed of countless individuals, each of them complicated and unique, interacting in massive and unimaginably complex systems, networks, and institutions in the context of constantly-changing information, norms, and technologies.

“In putting together our public opinions,” writes Lippmann,

“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.”

“The real environment is altogether too big, too complex, and too fleeting for direct acquaintance. We are not equipped to deal with so much subtlety, so much variety, so many permutations and combinations.”

There’s a tendency in thinking about democracy to ask whether voters are well-informed about “the facts.” For Lippmann, it’s important to reframe the issue in terms of a complex environment to emphasise that knowledge of narrow matters of fact is insufficient for understanding how a vast, complex reality is put together and how it works. A database of “facts”—even one as big as the human brain can store—is insufficient.

It’s also important for Lippmann to emphasise that the vastness and complexity of reality is not something people instinctively appreciate. He repeatedly notes that there is a strong but delusional tendency—what psychologists today call “naive realism”—to treat the truth about political issues as if it is self-evident to all well-intentioned observers.

Lippmann argues—rightly, I think—that this is wrong. The nature and structure of the modern world, and the causal mechanisms, forces, and laws governing its behaviour, are extremely complex and counterintuitive.

“The truth about distant or complex matters is not self-evident.”

—Walter Lippmann, Public Opinion

V. Constructing the pseudo-environment

Even if one grants that reality—the real environment—is vast and complex, why should this make real understanding impossible? That is, why must the pseudo-environments we construct inevitably simplify and distort the real environment in ways that undermine knowledge and good decisions?

Here, Lippmann advances several connected arguments.

Evolutionary mismatch

One theme that runs through the book is that if you step back and reflect on our contemporary condition, there is no reason to expect that knowledge or understanding will be possible. “Man is no Aristotelian god contemplating all existence at one glance,” writes Lippmann. “He is the creature of an evolution who can just about span a sufficient portion of reality to manage his survival.”

In other words, humans did not evolve to understand unimaginably vast, complex, large-scale societies. In a nice summary of Lippmann’s views here, the philosopher John Dewey (an admirer and critic of Lippmann) writes,

“The local face-to-face community has been invaded by forces so vast, so remote in initiation, so far-reaching in scope and so complexly indirect in operation that they are, from the standpoint of the members of local social units, unknown. . . . They act at a great distance in ways invisible to [them]”.

In other words, Lippmann is pointing to a kind of evolutionary mismatch. Our epistemic capacities—that is, our abilities to know and understand the world—are adapted to small-scale social worlds profoundly different from the modern world. Because of this, we mistakenly treat the modern world as more legible, more understandable, and more manageable than it really is.

Social mediation

In addition, Lippmann argues that our understanding of the real environment is almost entirely mediated to us via the reports of others. “The world that we have to deal with politically,” he writes, “is out of reach, out of sight, out of mind.” In other words, we do not confront reality “directly”. We construct our pseudo-environments largely from information that is second-hand, third-hand, fourth-hand, and so on.

In itself, it’s not clear why this should be a problem. Why can’t journalists just report on “the facts”, for example?

Lippmann thinks this is naive. For one thing, the number of facts is infinite, as is the number of ways of framing them, organising their presentation, situating them in context, and so on. In other words, even if journalists are solely concerned with reporting the facts, they are inevitably—that is, through no fault of their own—highly selective in which facts they report on and how they report on them.

In addition, journalists are in exactly the same position as the citizens they aim to inform. That is, a journalist’s own pseudo-environment is also based on information they’ve acquired from others, relayed to them via processes that are extremely selective by journalists with a similarly partial and indirect access to reality.

More importantly, just like citizens, journalists’ pseudo-environments are organised around what Lippmann calls “stereotypes”, simplifying interpretations and frameworks that render an unimaginably complex reality imaginable.

Stereotypes

Although the term “stereotype” as we use it today originates in Lippmann’s work, its current meaning is different from Lippmann’s. For him, stereotypes constitute the basic interpretative frameworks—the concepts, explanatory models, and principles—we use to organise information. In this sense, they’re not inherently “prejudiced”. Moreover, they’re unavoidable. As he puts it, the alternative to stereotyping would be “direct exposure to the ebb and flow of sensation”, a kind of impossible contact with reality unmediated by interpretation.

Stereotypes take a vast, differentiated, dynamic, ambiguous, and inaccessible world, one governed by subtle interactions between largely unknown forces and causes, and reduce it to a comforting, low-resolution mental picture of that world. In this process, the real environment is inevitably simplified and distorted in ways we’re not aware of.

Moreover, we cannot be aware of the misalignment between our mental pictures and reality. For any citizen, the pseudo-environment is the environment. When we compare our beliefs against the facts, we always find a comforting 1:1 correspondence. We can acknowledge the possibility of error in the abstract, but what these errors might be is an unknown unknown.

Importantly, Lippmann does not argue—implausibly—that pseudo-environments are simply delusions with no basis in reality. Instead, he argues that they are constructed in ways that are genuinely sensitive to facts, and they can incorporate corrective information. However, the order that they bring to experience comes at an inevitable cost: information which is congruent with the categories, explanatory models, and stories used to understand the world can be attended to. Otherwise it’s screened out.

It’s easy to see what Lippmann has in mind here if one compares the outlooks of explicit ideologues such as socialists and libertarians. In such cases, their worldviews are obviously not completely divorced from genuine facts and regularities. However, they organise information in fundamentally different ways—around different core beliefs (e.g., about human “nature”), simplifying categories (e.g., about “capitalism”, “the economy”, “neoliberalism”, etc.), explanatory principles (e.g., about markets, exploitation, etc.), and so on—which provide the background frameworks against which they judge the salience and importance of new information.

Lippmann’s point is that this is not restricted to ideologues. It’s an inevitable feature of all political cognition, including the implicit interpretative frameworks of citizens viewed—either by themselves or by political scientists—as “non-ideological”.

Moreover, whereas people tend to think that some stereotypes—their own—are right and some are wrong, Lippmann thinks this is simplistic. In an important sense, they’re all wrong. More specifically, they all constitute gross simplifications and distortions of a complex reality, attending to genuine features of that reality but in ways that are extremely partial and distorted.

All of this is nicely captured in the following quote from The Phantom Public:

There is a bias in all opinion, even in opinion purged of desire, for the man who holds the opinion must stand at some point in space and time and can see not the whole world but only the world as seen from that point. So men learned that they saw a little through their own eyes, and much more through reports of what other men had seen. They were made to understand that all human eyes have habits of vision, which are often stereotyped, which always throw facts into a perspective; and that the whole of experience is more sophisticated than the naïve mind suspects.

VI. A technocracy?

If democracy depends on forms of knowledge and understanding that voters can’t achieve, what is the solution? At the end of Public Opinion, Lippmann recommends replacing democracy as it is conventionally understood with a kind of technocracy, a political system in which the job of understanding and solving problems is outsourced to a class of technical experts. By applying rigorous scientific and statistical procedures and building upon well-established data and expert knowledge, these experts can overcome the epistemic limitations of ordinary citizens.

Lippmann doesn’t say much about how this is supposed to work. Reading Public Opinion, I get the sense he felt the need to offer some solution to the problems he identifies, but didn’t devote much thought to the topic. However, it’s clear that his proposal rests on a core assumption: that the sources of error and fallibility in the pseudo-environments of ordinary voters are not inevitable features of political cognition. And underlying this assumption, I think, is his view that rigorous scientific and statistical methodologies are both (i) alien to how human beings instinctively think about the world and (ii) capable of generating genuine knowledge.

VII. Assessment

It’s been over a century since Public Opinion. It’s plausible that we now know a lot more about the world, human psychology, scientific methodology, and political epistemology. From this vantage point, I think it’s clear that Lippmann was wrong about many things but deeply insightful about others.

What Lippmann got wrong

1. All models are wrong, but some are useful.

Ironically, one very basic problem with Lippmann’s argument is that it rests on a model of political epistemology that is too low-resolution. He’s right that citizens are forced to rely on mental pictures or “pseudo-environments” that in a deep sense are always wrong. However, as the famous aphorism has it, “All models are wrong, but some are useful.”

Understanding comes in degrees. If your pseudo-environment depicts the government as Satan-worshipping lizards, there’s nothing positive to say about it. In contrast, if you take time to acquire information about the world from reliable sources and process the information in reasonable ways, your mental picture of reality will no doubt be incredibly simplistic and incomplete, but it can have significant epistemic value nevertheless. At times, Lippmann’s arguments conflate a plausible fallibilism (the truth is not self-evident and worldviews are inevitably partial and mistaken) with an implausible scepticism in which there’s no reason to believe any mental models are superior to others.

2. How you think and reason matters

Relatedly, Lippmann’s arguments don’t do justice to profound differences in the ways in which citizens seek out and process political information. Although there’s a sense in which all political cognition rests on simplifying stereotypes, work by Philip Tetlock and others has shown that there are consequential differences in the degree to which people are “ideological” in how they form their beliefs.

Comparing the ability of people to forecast events, Tetlock has found that bad forecasters tend to exhibit recurrent traits:

“As ideologically diverse as they were, they were united by the fact that their thinking was so ideological. They sought to squeeze complex problems into their preferred cause-effect templates and treated what did not fit as irrelevant distractions. Allergic to wishy-washy answers, they kept pushing their analyses to the limit, using terms like “furthermore” and “moreover” while piling up reasons why they were right and others wrong. As a result, they were unusually confident and likelier to declare things “impossible” or “certain.””

There’s no doubt that this characterises how many people—probably most people—approach politics. In this sense Lippmann’s analysis of stereotypes was insightful. However, Tetlock has also shown that superforecasters—that is, those capable of predicting events with at least some (highly limited but real) success—tend to exhibit very different traits like intellectual humility, the acknowledgement of uncertainty, a willingness to engage with and integrate diverse perspectives, genuine curiosity, and most of all an ability to treat beliefs as “hypotheses to be tested, not treasures to be guarded.”

Citizens who think in this way are genuinely different from ideologues and capable of outcompeting them in their understanding of the world.

3. Social epistemology

Although Lippmann’s work is a profound treatment of political epistemology, there’s an important sense in which its understanding of epistemology is still highly individualistic. Specifically, it focuses almost entirely on defects in individuals’ mental pictures of the world. However, human beings are epistemically cooperative, and groups of individuals can exhibit characteristics that give rise to collective forms of knowledge far beyond the reach of individuals. For example, there’s the wisdom of crowds, the benefits of epistemic communities characterised by intellectual and cognitive diversity, and status games which reward those who improve the epistemic commons.

Admittedly, it’s an incredibly difficult question whether democracies really benefit from social-epistemic processes like this, but it’s noteworthy that Lippmann doesn’t really engage with the question at all.

4. Journalism

Lippmann’s analysis of journalism is also far too coarse-grained. He’s absolutely right that journalists must be incredibly selective in which facts, events, and trends they report and in how they report on them. However, there are consequential differences in the quality of different media outlets here that his analysis neglects.

In recent years, for example, there has been a push by some journalists and media outlets to report on the world in ways that convey reliable knowledge about statistical trends, which is both possible and desirable. There really is a difference between media outlets that report on a highly non-random sample of all the bad things happening in the world, which systematically distorts audience perceptions of real trends, and organisations like Our World in Data that strive to inform audiences about genuine regularities and patterns rather than attention-grabbing anecdotes.

5. Expert Limitations

Lippmann’s advocacy of technocracy (i.e., rule by experts) is often criticised as being “elitist.” This isn’t a good criticism. Knowledge, talents, abilities, skills, and so on aren’t equally distributed among people. Elitism is often good. Moreover, I think Lippmann was right that well-trained, knowledgeable experts are in principle often capable of outcompeting ordinary voters in acquiring knowledge about the world.

However, Lippmann was still far too optimistic about the epistemic capacities of experts. Just as work by Tetlock has shown that genuine understanding of even complex trends is at least to some degree possible, it has also shown that much expertise is fake. Moreover, the past decade has revealed that much of what passes for social-scientific knowledge is in fact extremely unreliable. Certainly when I think about the areas of social science I know best—for example, “misinformation studies”—it’s clear to me that “experts” have not acquired substantial knowledge of a sort that can solve the problems that concern them.

Some of these issues can be addressed by better scientific practices and methodologies. However, there are profound limitations on what experts can achieve even in principle, limitations which arise for reasons that Lippmann identified: namely, the vastness, complexity, and inaccessibility of reality.

When it comes to modern societies, careful randomised experiments are almost never possible, and even when they are possible it’s very unclear whether and how their results generalise to non-experimental contexts. When it comes to non-experimental social science trying to discover causal knowledge from statistical relationships, I think expert scientific “knowledge” is basically fake and actively harmful because it misleads people into thinking they know things when they don’t.

6. Justifying democracy

Finally, Lippmann assumes that democracy only “works” if citizens are capable of becoming sufficiently well-informed to make good decisions. But this is extremely dubious. Even if citizens are outright delusional and democracies make many bad decisions, democracy still might be better than alternative political systems. Moreover, democracies might have many desirable properties that Lippmann doesn’t explore. For example, they symbolically convey “equality” of a certain kind, which many people care about and which might reduce the risk of political revolutions and violence. Further, they might also help to change who holds power more frequently than other political systems, and even if they don’t promote good decisions, they might be effective at guarding against extremely bad ones.

Importantly, I’m not saying I think democracies do have these virtues. But they might, and Lippmann doesn’t really explore their possibility.

What Lippmann got right

Despite these objections, I think Lippmann nevertheless gets a few very big things right.

First, we all tend to greatly exaggerate the degree to which we understand the modern world. Even if understanding comes in degrees, and some citizens and experts do seem to acquire something recognisable as genuine knowledge, there are profound limits on understanding that derive from the kinds of factors Lippmann identified. To a greater extent than most of us are willing to acknowledge, our perspectives on the world are simplistic and partial, and reality is mostly unpredictable, unknowable, and unmanageable.

Second, this would still be true even if you eliminated all the distorting factors that corrupt political judgement: for example, our motivations to signal tribal allegiances and behave as our own press secretaries, interpreting information in ways that are self-serving and self-justifying; and the role of political, economic, and cultural elites in spreading self-serving lies and bullshit through media and other sources.

Finally, Lippmann is right that our naive realism is profoundly harmful. Because we instinctively treat the truth—even about complex political matters—as self-evident, we greatly over-estimate our abilities to shape the world to our desires. Moreover, if the truth is self-evident, there must be something wrong with those who fail to acknowledge the truth. They must be liars, victims of lies, or insane. Even setting aside bias and tribalism, this seems to shape how many people approach political disagreement in ways that generate unnecessary hostility and conflict.

For these reasons alone, Public Opinion is one of the most important works of political theory ever written. It’s also one of the most underrated.

Further Reading

• I would highly recommend Jeffrey Friedman’s brilliant book Power without Knowledge for an incredibly insightful and radical development of Walter Lippmann’s views.