The following is an excerpt from Why It’s OK to Speak Your Mind, by Hrishikesh Joshi. Routledge, 196 pages. (March 2021)

The division of cognitive labor

Modern society is only possible because of the division of labor. Without division of labor, the most we could achieve is a very meager standard of living. Imagine you had to make everything you use, by yourself, from scratch—without tools created by others, without water and food provided by others, without medicines invented by others. Most of us would not survive for a month, if that. Division of labor makes modern standards of living possible because with individuals specializing in one area, society as a whole is able to be much more productive.

Adam Smith illustrated and developed this idea in his Wealth of Nations by using the example of a pin factory. Imagine 10 people tasked with making pins. If each person had to make a whole pin, perhaps each might make 10 pins a day. Making a whole pin involves several distinct processes. Let’s suppose it involves 10 different tasks. Well, if one person had to do all these tasks we can expect that there would be time lost as that person transitioned from one task to another. Furthermore, it would be hard to become skilled at all these different tasks—that would require lots of training and effort. But what if each person in the factory focused on just one of the 10 tasks instead? Time would be saved in a myriad of ways, and the factory would be able to produce a lot more pins—though, no person by himself would be making a whole pin. As a result, the factory might produce 10,000 pins total per day, whereas it would have produced only 100 without specialization. Modern society is like this pin factory writ large.¹

But division of labor in modern life is not limited to the production of physical goods. The other face of specialization is the division of cognitive labor. Our institutions of knowledge production (universities, thinktanks, private research labs) reflect this feature: researchers inevitably specialize in one tiny sub-subfield or two in order to make new discoveries. Yet, the division of cognitive labor has deep implications. What we are able to know is inextricably tied to what I will call the epistemic commons—the stock of facts, ideas, and perspectives that are alive in society’s discourse.

In their book, The Knowledge Illusion, cognitive scientists Steven Sloman and Philip Fernbach write: “Language, memory, attention—indeed, all mental functions—can be thought of as operating in a way that is distributed across a community according to a division of cognitive labor.”² The authors argue that we know very little, but take ourselves to know a lot because the relevant facts are easily accessible to us. If Sloman and Fernbach are right, then our epistemic health as individuals—i.e., the extent to which our beliefs accurately represent the world—is inextricably tied to the health of the epistemic commons.

Consider the following. Do you understand how a zipper works? How about a flush toilet? These objects seem basic enough. Knowing how they work isn’t exactly rocket science. But people drastically overestimate their understanding of how these simple items function. In one study, Leon Rozenblit and Frank Keil asked people to rate from one to seven how well they understood the workings of such objects. They then asked participants to actually explain in detail how the objects worked. Many were simply unable to do so. And so when asked to revisit the question of how well they understood, subjects drastically lowered their ratings. Psychologist Rebecca Lawson performed a similar experiment where students were asked to explain, by sketching out the mechanism, how a bicycle works. The results were striking—most people were unable to complete the task, even though a bicycle is such a familiar object in our daily lives. This phenomenon, of people thinking they know much more than they actually do, has come to be known as the illusion of explanatory depth.³

Why might we fall prey to this illusion? Well for one, the relevant information is easily accessible. If you want to know how a zipper really works, a simple Internet search will give you all the details you need. Though you may not actually as of this moment know the workings of a zipper, the knowledge is “at your fingertips,” as it were. What this suggests is that our representation of the world is like a low-resolution map such that “zooming in” only gives a clear picture insofar as we are able to rely upon the knowledge others have. With respect to most areas of the map, we are unable to zoom in by ourselves—and if we do, we’ll just see large pixels that don’t look like anything. The division of cognitive labor, then, renders our epistemic lives intricately tied with the efforts and contributions of others.

Furthermore, the very coarse-grained picture we have of the world will itself depend on which perspectives are “alive” in the discourse within our milieu. Consider for example a teenager within a deeply religious sect living in a small village. Suppose that this sect does not believe in Darwinian evolution. The arguments for evolution are not discussed, and when the topic is broached, people quickly dismiss it as an unsubstantiated theory. Some might raise what they take to be decisive counterarguments like: “How come we don’t see monkeys turning into humans now?” or “Where are the missing links?” and so on. Now the teenager might be able, in principle, to discover the powerful arguments in favor of evolution by natural selection. There is a copy of the Origin of Species at the local library, and she could also spend time delving into encyclopedias and biology textbooks. But for all intents and purposes, her map of the world has a large hole in it. What’s more, given that there are ample other constraints on her time, she might simply not find it worthwhile inquiring further.

In this way, there are lots of questions that we might lack the time or imagination to inquire about if the people we’re surrounded by consider the issue settled. Division of cognitive labor means we simply cannot independently verify all the claims we take for granted. But that in turn means that if the view our community settles on is mistaken or impoverished, the distortion easily transfers to us. Our epistemic health thus depends on the epistemic health of our milieu.

The 19th century mathematician and philosopher W.K. Clifford underscored this social, interconnected nature of our ability to understand and describe the world in his landmark essay on the ethics of belief:

Our lives are guided by that general conception of the course of things which has been created by society for social purposes. Our words, our phrases, our forms and processes and modes of thought, are common property, fashioned and perfected from age to age; an heirloom which every succeeding generation inherits as a precious deposit and a sacred trust to be handled on to the next one, not unchanged but enlarged and purified, with some clear marks of its proper handiwork. Into this, for good or ill, is woven every belief of every man who has speech of his fellows. An awful privilege, and an awful responsibility, that we should help to create the world in which posterity will live.⁴

For Clifford, this meant that each of us has an important ethical responsibility: namely, to believe only on the basis of proper evidence. As I will be arguing in the next chapter, if our epistemic situation is a common resource in this way, then we all have a duty to do what we can to preserve the integrity of this resource. However, believing on the basis of proper evidence, though important in its own right, is not enough—we also have a duty to speak our minds.

Blind spots and social pressure

To set the stage for that argument, it is necessary to examine the way in which the epistemic commons is vulnerable. Tragedies of the commons arise because common resources are often susceptible to damage and degradation.⁵ For example, industrial pollution can destroy river ecosystems. Analogously, I will argue below, social pressure can degrade the epistemic commons.

Consider again the village described above. Why might reasons to accept evolution be systematically repressed here? Presumably because publicly defending such reasons will come at some cost to one’s social status, the maintenance of which is a strong motivation for most people.⁶ Somebody discussing evidence in favor of Darwinian evolution might be seen as deviant, and perhaps not a true believer of the religion. Furthermore, accusations of heresy or disbelief can invite severe repercussions in many deeply religious societies—even if such accusations end up being untrue. Thus, even if somebody were to encounter or think of a reason to believe in evolution, they might keep that thought to themselves, especially if they’re unsure of the soundness of the reason. Why risk your reputation and social standing (or worse, in many places and times) just to voice some reason you’re unsure of?

In this way, social pressure can systematically filter out reasons to believe a particular claim. The reasons that don’t get filtered out will make it look like that claim ought to be rejected—even if had there been no such filtering, then people would be justified in believing the claim. In other words, filtering processes created by social pressure allow reasons to pile up on one side of an argument while those on the other side get discarded. Yet, the overall balance of reasons, had open discourse prevailed, might well have supported the other side. Any time we observe social pressure to avoid giving some kinds of reasons, then, we should suspect that a worrisome blind spot exists in some form or another.

Importantly, we can’t dismiss the existence of such distortions simply by surveying the first-order evidence (i.e., evidence directly relevant to the issue at hand) presented to us. The problem is created precisely because evidence is filtered in such a way as to support one conclusion. It’s then no good to simply look at the evidence that is presented and say: “but the conclusion is obviously right!” The conclusion looks obviously right because countervailing evidence is not allowed to surface and accumulate, due to the presence of social pressure. A collective blind spot can exist in this way even if the members of a community respond rationally to the first-order evidence they have.

Lessons from the 20th century

Evidential situations like these can lead to catastrophe. If information is not freely shared within a group due to social pressure, deliberation on very important issues can be distorted. In the above case, the dam will break and ruin many lives.

Moreover, this is not simply an exercise of the imagination. Many avoidable disasters have occurred because there was pressure not to share certain kinds of information. The Chernobyl disaster, in which a nuclear powerplant malfunctioned and exploded in what is modern-day Ukraine, is perhaps a paradigm example. Due to the authoritarian, top-down government in place at the time, individuals had incentives not to raise alarms about radiation levels, the nature of the explosion, substandard materials, etc. The result was devastating for thousands of people, many of whom continue to feel the effects of radiation poisoning to this day. The HBO series Chernobyl offers a detailed look at the deliberations and actions of various individuals as they grappled with the situation in a way that brings out the incentives they had to distort or suppress information.⁷

Democracies typically do a better job of avoiding unnecessary disasters and missteps like this. The victory of the Allies in World War II can be partly attributed to the nature of information flow within democratic decision-making.⁸ In the democracies, members of the army were relatively more able and willing to offer information that would lead to course-correction by the upper chain of command. By contrast, within the German army and air force, people were much more hesitant about displeasing their superior officers with news or information or strategic perspectives that might be seen to dampen the war effort.

Democracies are also able to allow the spread of key information through a more open media. Journalists are less prone to intimidation by the government, and thus can quickly disseminate crucial news to civilians and government officials alike. Luther Gulick, who served as a high-level American official during World War II, explained that in contrast, decisions within authoritarian governments are “hatched in secret by a small group of partially informed men and then enforced through dictatorial authority.”⁹ Democracies are thus able to avoid some of the epistemic pitfalls that beset authoritarian regimes because the channels of information are much freer.

This is no cause for complacency, however. Democracies are not immune to such problems. For example, the infamous Bay of Pigs Fiasco, a failed US-backed landing attempt on Cuba in 1961, resulted in part because those who had doubts about the plan suppressed their reservations.¹⁰ Moreover, social pressure need not always come from government authorities. Think of college students who feel pressure to binge drink, the many of us who feel pressure to dress in particular ways, teenagers who (used to) feel pressure to smoke cigarettes—or, what’s more relevant here, people who feel pressure not to publicly express certain social or political opinions. Such forms of social pressure do not come top-down, from some governmental chain of command. Rather, they are much more spontaneous and organic. These pressures emerge from the incentives, interactions, and choices of millions of people who shape a particular culture. Democracy, then, does not solve all the informational problems systemic within authoritarian regimes.

The danger today

In her groundbreaking work on the dynamics of public opinion, political scientist Elisabeth Noelle-Neumann argued that fear of isolation can create a “spiral of silence,” where only one side of an issue is publicly defended. The core mechanism she identifies is this. People don’t want to say things that they believe might risk eliciting the disapproval of others; they don’t want to potentially lose friends and get pushed out of their social groups. There is a fear of isolation. So, instead of saying what they really think about a particular issue, such individuals keep mum. Once the process is set in motion, more and more people become silent about their true opinions.¹¹

Spirals like these typically occur with regards to contentious, emotionally laden moral and political issues. A spiral of silence can drive even the majority opinion underground if the minority is sufficiently vocal, and especially if mass media repeatedly and concordantly come down on one side of the issue. Eventually, the spiral of silence causes the majority opinion to effectively disappear, while the previously minority opinion becomes the dominant societal assumption.¹²

What does this mean for us, now? Well for one, we shouldn’t assume, for all the reasons explored so far, that such spirals of silence induced by social pressure (real or perceived) are going to line up with the truth all the time (or even most of the time). Spirals of silence are sensitive to social forces, not to the truth. Thus, they can cause society to settle on opinions that are quite misguided.

However, in order to know what policies to support or how to remedy various social problems, we need to have an accurate idea of what the social world is like. The very best of intentions can have terrible consequences if those intentions are not supplemented with an accurate picture of the world. (Indeed, under some description, more or less all of the worst actors and movements in history can be said to have “good intentions.”) But social pressure can warp our collective picture of the world without individuals being in good positions to detect the distortion. So, the more we allow spirals of silence to occur, the more chance there is for the road to hell to be paved with good intentions.

The danger we face today is that many of us have quite confident views about lots of contentious issues, as well as lots of issues that have been “settled,” not via a process of institutionalized disconfirmation, but rather through spirals of silence. But this means that the steps we might take to mitigate economic and social problems could backfire, making things worse. The risk becomes greater the more radical, as opposed to piecemeal, solutions we embrace. We might also be misdiagnosing what the problems are in the first place. And we might be missing various forests for the trees. Our Chernobyl, so to speak, might not involve a nuclear powerplant, but might instead manifest itself in the way we conceive of and try to solve social and economic problems.

One way to respond to this predicament is to encourage epistemic humility.¹³ Perhaps we should all just check ourselves. This however, is far easier said than done. Knowing our epistemic limitations in abstract terms may not actually induce humility in us (especially the loudest among us) when the rubber meets the road. The only way to properly mitigate our dangerous blind spots is for courageous individuals to speak their minds, and refuse to buckle to social pressure. This is not to say that epistemic humility and other tools for critical thinking are not important or worth cultivating. But if knowledge is a collective enterprise, individual epistemic humility can only go so far. This humility, for instance, cannot prevent a Chernobyl—only people sharing their evidence can.

Institutions of knowledge production

Social pressure creates blind spots by making it costly to provide evidence on one side of an issue, while making it costless or even beneficial to provide evidence on the other side of the issue. Whenever such incentives exist, we should suspect that our resulting view of the world is warped in some way. These incentives are particularly important to address within the institutions responsible for knowledge production and dissemination: research groups of various sorts and fields and academic departments within the university system.

Given modern division of labor, such institutions specialize in knowledge production; the rest of society thus relies upon them for providing an accurate picture of the world. Other individuals in society, however, do not have the time or resources to check all the work produced by such institutions, and so an element of trust is necessary. Analogously, you don’t have the time or wherewithal to check all the work done by your lawyer, doctor, or accountant—when it comes to your interaction with such specialists, then, an element of trust is involved.

However, social pressures within institutions responsible for knowledge production can undermine their mission and distort their product. Science works well only in a context of institutionalized disconfirmation: that is, a situation wherein researchers are free and even incentivized to disconfirm any and all hypotheses that are in contention.

Over time, science has disconfirmed hypotheses that would seem exceedingly natural to humans observing their world. Many things that seem intuitive to us turned out to be false. The Earth, it turned out, is roughly spherical, though it looks flat from our vantage point. And while the sun looks like it goes around the Earth, the reverse is true. In the 17th century, Galileo Galilei suffered persecution at the hands of the Catholic Church for defending this idea. Science naturally works best when such costs are absent—so that it doesn’t take a Galilean personality to seek the truth.¹⁴

Modern physics has upended our intuitive picture of the world even further. The things that look “solid” to us—tables, rocks, books, etc.—are actually made mostly of empty space.¹⁵ And the fundamental units of physical reality have both particle-like and wavelike properties. Albert Einstein famously showed that time is not absolute. Whether or not two spatially distant events are simultaneous depends on the observer’s frame of reference. He further showed that space and time are intertwined in such a way that it’s best to think of them as spacetime. According to the best models we currently have to explain the behavior of large objects, gravity is the result of spacetime “bending” around massive objects.¹⁶ Trippy stuff!

How has science made these remarkable discoveries that are so far from our intuitive sense of the world? Science is a collaborative effort, and no one person can do it all by themselves, even within a sub-sub-field. Science involves enormous division of labor. But for us to be able to trust the products of science, the incentives have to be right. The incentives that individual scientists face must be aligned with finding the truth, wherever it may lie. Generally, this is the case, and that is why science has been on the whole very successful. In physics or chemistry, if you are able to find experimental data that disconfirms an important and commonly accepted hypothesis, you will receive many professional goods—you’ll likely get published in prestigious journals like Nature or Science, you might get big grants in the future, an endowed chair, maybe even the Nobel Prize.

Given these incentives, physics and chemistry are self-correcting. If a hypothesis is easily disconfirmed, it won’t last for long. Researchers, incentivized to disconfirm it, will quickly design experiments to show why the hypothesis doesn’t hold. Sloman and Fernbach write: “Scientific claims can be checked. If scientists are not telling the truth about a result or if they make a mistake, eventually they are likely to be found out because, if the issue is important enough, someone will try and fail to replicate their result.”¹⁷ Many scientists have echoed the importance of this feature of science over the years. Any time the accepted wisdom strays from the truth then, a course-correction will quickly follow.

Understanding knowledge production as a collective endeavor, which relies heavily on a well-maintained epistemic commons, helps us appreciate why John Stuart Mill defended his somewhat radical sounding account of justification for our scientific beliefs in On Liberty. He wrote:

If even the Newtonian philosophy were not permitted to be questioned, mankind could not feel as complete assurance of its truth as they now do. The beliefs which we have most warrant for, have no safeguard to rest on, but a standing invitation to the whole world to prove them unfounded.¹⁸

Thus, imagine if critics of Newton’s physics found themselves unemployable or prone to receiving censure, threats, etc., as soon as they challenged part of the view. Could a person living in Mill’s time, circa the mid-19th century, be able to trust the science of physics? Could he have reasonably believed in Newton’s laws if people faced a very uphill battle in trying to disconfirm them and he knew about this situation? Plausibly not. For, especially if this person is not a physicist, he lacks the wherewithal to check the researchers’ work. For all he knows there may be good reasons to reject Newtonian physics that are just not allowed to surface.

Indeed, as it turns out, Newtonian physics was accurate only in approximation. For macroscopic objects traveling at relatively low speeds, i.e., well below the speed of light, Newton’s laws allow us to make approximately true predictions. However, as Einstein later showed, some decades after Mill had passed away, Newtonian physics breaks down when it comes to objects moving close to the speed of light. Furthermore, while Newton assumed that space, time, and mass are absolute, Einstein showed that they are relative. Which events are simultaneous, how long an object is, how much mass it has, all depend on the observer’s frame of reference. If you are traveling at, say, half the speed of light relative to where I stand, then the length of a particular table will be quite different for you as opposed to me. Hence, even Newtonian physics, which was by Mill’s time well established and confirmed with countless experiments, turned out not to be sacrosanct.

The scientific process, then, must be structured in a certain way for it to merit our trust and reliance. If there were contrary evidence to be found, would it be discovered, published, and incorporated into the mainstream scientific consensus? The answer to this question must be yes.

In some sense, the scientific enterprise must be objective. What does such objectivity mean? Philosopher Helen Longino argues that it requires an openness to what she calls transformative critique. For Longino, science is fundamentally a social practice, and it is precisely due to this fact that its objectivity can be secured. Individual researchers are bound to have their idiosyncratic perspectives and biases. However, “science” is not simply the aggregation of the findings of individual scientists. Science is fundamentally practiced by social groups, not lone individuals. What gets counted as scientific knowledge results from social processes like peer review, attempts at replication, citation patterns, and clashes between defenders of alternative hypotheses and paradigms. This is a feature, not a bug. “Only if the products of inquiry are understood to be formed by the kind of critical discussion that is possible among a plurality of individuals about a commonly accessible phenomenon,” says Longino, “can we see how they count as knowledge rather than opinion.”¹⁹ Consequently, the more diverse points of view there are within a scientific community, the more objective the process is likely to be.