Trust and consensus




Post-truth often occurs when knowledge is hard to find: difficulties are amplified and exaggerated and confusion, lies and concealment spread. In factual matters, for us to tell what is true from what is not (always in practical terms) we must understand not only whether there is evidence, and of what quality, broadly speaking, but also where the weight of evidence lies. 

This body of evidence that we, as humanity, are gathering is a body of knowledge constantly being expanded and remodeled. Evidence does not always point in the same direction, which is only natural: there may be methodological or interpretation errors, instruments may improve or fail, or there may even be some natural variation in the facts. But if the vast majority of experiments and observations give similar results, that is where the weight of evidence lies. 

Today, not only the speed with which new knowledge appears is increasing, but also the way it is conveyed to society. This means that all of us have that knowledge, or at least a description of that knowledge, at our fingertips with increasing ease. At first glance, it might seem that this protects us against post-truth: if we have more information, then it is easier to find out whether a given statement is correct or not. Is this really the case? Not really, and perhaps the exact opposite is true. 

Besides the fact that there may be contradictory evidence, there may also be self-interested influences, especially on issues that resonate more with politics or society. In these situations, it is not easy to decide what can and cannot be trusted. If we add that difficulty to our own cognitive biases,1More on this in chapter VI. it is also increasingly easy to convert our previous position on an issue, consciously or not, into one that is seemingly evidence-based. Thus, we are faced with one of the most treacherous (to others and to ourselves) forms of information management: instead of evidence-based decisions, we have decision-based evidence: it is very easy to select, from the total body of available evidence on a topic, that subset which best aligns with our position. 

Knowledge is more than a mere collection of facts from which we can select those we like. In this case, since we have some evidence, we believe more firmly that our judgment is correct, which makes the quantity and actual availability of evidence a double-edged sword. There's something not quite right about this, isn't there? There is such an abundance of evidence today that in the end it is becoming easier, not harder, to surrender to post-truth if we are not careful. 

If 100 different observers say something and 2 say the opposite, where does the weight of evidence fall? In principle, it would seem that in the first group. But this is only true if the 100 observers are independent, that is, if they are not related to each other and each one addresses the same problem with their own methodology. On the other hand, if they are merely dissemination channels that repeat what originated from a single source, they have no additional value, and what they do is generate an illusion of consensus and knowledge that is not so. 

We need, then, to distinguish real knowledge from the illusion of knowledge; the voices from the echo. It is not enough to get some isolated evidence to support our ideas. 

And this is also true for this book --or, at least, that is my intention. If I tell a story or an anecdote, it is to illustrate something that has been mostly validated and not to justify it. If something doesn't have much supporting evidence, or if I don't know if it does or not, I say what I know. If I don't have any evidence, or hold something as a personal opinion, I also try to make that clear. 

If we want to survive the traps that post-truth puts in our way, we have to be more discriminating. How do we analyze knowledge? How do we use it? Which statements can we trust? And, above all, to what degree can we trust them? This is not easy, because we have to find a delicate balance between extremes that do not help us: total trust and total distrust.2The important word in this sentence is fair. It is not taking any two opinions and averaging them. You have to understand the evidence and its qualities in order to weigh them fairly.

To find out what something is like, to get closer to knowing that truth, we had said that we can look for and interpret evidence obtained through observations and experiments. How much to trust or not to trust each piece of evidence is a more sensitive matter, but certain general rules can be established that, although not infallible, serve as a guide, as we mentioned in the chapter on medicine. Now, we will complement that with a new approach, a new layer of complexity, which can be more general as it applies more clearly beyond scientific and medical issues. 

Let's start, then, with simpler situations. Often, most experts in a particular subject agree, broadly speaking, on the central ideas in their field. Even if they disagree on some points, these are details. That basic agreement is reached because they assess the total body of available evidence in terms of its quality and quantity, and see that it all points in the same direction, regardless of whether any particular piece of evidence contradicts it. When the evidence is multiple, independent, comes from different fields and converges in a single, coherent interpretation, we speak of convergence or concordance of evidence. This is something very powerful and not at all easy to achieve. As we said before, it is important that these experts generate and assess the evidence independently and not by echoing each other. In these cases, we speak of scientific consensus

In defense of independence, there is even some evidence3See Danchev, V., Rzhetsky, A. and Evans, J. A. (2018). Centralized big science communities more likely generate non-replicable results. Available online. that the collective correctness of science, the way in which every claim is tested, in principle, permanently, works much worse among scientists who are connected (i.e., who collaborate with each other) than among those who are not. This seems to support the idea that, in order to produce robust results, it is important that scientific activity is not inbred, that there are research groups that address problems independently, and that this is facilitated by scientific policy decision makers. 

Scientific consensus is very different from consensus in our daily or political life. As we mentioned before in connection with the word theory, we must bear in mind that technical terms sometimes have a particular meaning within a field and a different meaning outside it. Something similar happens with the word consensus. In everyday language, we use it for situations in which a group agrees on something. Perhaps there are individuals in the group who disagree but, due to particular situations, decide that, in spite of that, they will support the group in the direction it takes. In these cases, the agreements may be motivated by social issues, politics, etc., and not necessarily by the search for the truth or the best possible solution to something. In general, in the common sense of the word, the idea of consensus points to that of negotiation. If we go out to eat with friends and some want to have Chinese food and others want barbecue, maybe we end up going to eat barbecue at a Korean restaurant: we reach a consensus on something that is satisfactory for all of us, even if it is not what we originally wanted. But scientific consensus is something else. When there are issues that in practice are already considered settled, we can say that a scientific consensus has been reached. Here, the word consensus is understood as consensus on the interpretation of a broad, diverse and independent body of quality evidence, and not on what a group of scientists may think without evidence. 

For example, here I am doing what I said before. There is not much evidence about this. One piece of evidence that exists (and that I found, because of course there could be others I don't know about) was quoted so that anyone can look it up and read it. Another thing I do is to use "cautious language", such as "evidence pointing to" or "this seems to support the idea that", also as a way of showing the degree of certainty that the information seems to have.

There is no formal definition of when consensus is reached on an issue, and, of course, even when it is reached it is not static, but continues to change shape based on new evidence and interpretations. Consensus emerges gradually, and is always based on evidence and on how multiple different observers interpret that evidence: as quality research is added, approaching a problem from different points of view and with different methodologies, as the evidence supports a particular interpretation, consensus on a topic is generated. Sometimes, the consensus changes abruptly, when new evidence appears that shows what was missing in order to understand things differently. 

We saw that Linus Pauling, when he claimed that vitamin C prevents colds, was totally wrong. At that time, there was still no consensus because research had just begun, but today the scientific consensus is clear on this issue: vitamin C does not prevent colds. 

There are many things that scientific consensus is not. Perhaps, we imagine that scientists get together and, by means of a vote, define whether there is consensus or not on a topic, but it is not (really) so. Consensus is, broadly speaking, the outcome of several processes. First, the available evidence on a topic, including its interpretation, is openly and permanently debated. Based on that, multiple different professionals hold, with greater or lesser confidence, the idea that seems most accurate in terms of the supporting evidence. Thus, scientific consensus represents the aggregated interpretations of the available evidence by the experts on the subject. 

Even once the consensus has been established, there may be voices that oppose it, but these voices are generally few and usually come from people or organizations that are not experts in the subject of study, or that have some kind of interest in contradicting the consensus. 

We have already said that consensus is not written in stone, but for it to be modified, powerful, consistent evidence is needed. A consensus will not change because a group of people do not agree with the accepted explanation and argue "we do not agree, but we have not yet found the evidence to support what we say", or "we do not agree, and here is some supporting evidence". 

For example, there is consensus that the Earth is a spheroid, not a flat planet. Although there are people today who still maintain that the Earth is flat, no astrophysicist thinks it is. In the 4th century BC, the possibility of a flat Earth was conceivable (and even so, Eratosthenes measured its circumference in the 2nd century BC), but to continue to think so now goes against everything we know. It is not that the Earth has changed. What has changed is our knowledge of the world. 

Once consensus has been reached on a topic, is that it? Does the research or reinterpretation of previous evidence stop? What if a group of scientists does not agree with the consensus in their field of expertise and wants to challenge it? Perhaps they take into account evidence that was discarded by others, or have alternative interpretations. What happens in these cases? Generally, if someone wants to challenge the consensus, they will be asked for incontrovertible evidence to support their position. The stronger a consensus is, the greater the quantity and quality of evidence required to contradict it. If some want to say that the Earth is flat, it is not enough for them to say that there is a conspiracy going on, or that every photo of the planet taken from outside the atmosphere by satellites is a fraud. They should also be able to find alternative explanations that are really convincing and compatible with a flat Earth for the existence of day and night, seasons, airplane flights, time zones, meteorological phenomena and photos of the planet taken from outer space. 

Let's take another example: in recent decades, we have observed that global extreme poverty is gradually decreasing. Not only is there a smaller percentage of people living on less than $1.90 a day, but the absolute number of people in this situation is decreasing. In another example of "data kills intuition", the world may not be in as bad a shape as we sometimes think. 

That extreme poverty is decreasing is a consensus based on measurements and estimates, not on the unfounded opinion of some people. There may be some criticism about how to measure it, and even how to define it. Moreover, the fact that it is decreasing globally does not mean that it may not be increasing in some specific places. But, even considering all those differences that may appear among experts, there is agreement that the trend is that we observe a gradual decrease in extreme poverty. 

As we can see, there is scientific consensus on issues that are not necessarily identified as scientific: what’s scientific about them is the process that led to their establishment. In the case of poverty, these are observations made on the basis of very careful measurements. 

So far, we have seen situations in which the evidence is generous and the consensus is quite clear, and they involve issues that do not arouse strong emotional responses or do not directly contradict what benefits some powerful groups. But, of course, some situations are more complex.


Sometimes, we do not yet have a consensus, and perhaps we will not have one in the near future. This is often the case for topics where research is still very recent and the results are still ambiguous, or for topics that are extremely complex to address. For example, we do not know well how consciousness is generated in the brain (this is a case of an extremely complex and relatively new problem in research). Nor do we know how new media and algorithms impact the construction of societies. We suspect that they do so in a very dramatic way, but it may be that we are the dramatic ones. It is still very difficult to draw conclusions. 

Or maybe there is a consensus, but it is not very solid. Moreover, it is not categorical, but gradual. For example, in the field of nutrition, there is still no clear consensus on what can be considered a healthy diet. Yes, the broad ideas are: eat more fruits and vegetables, less processed foods and, in general, reduce the consumption of salt and sugar. Doubting this is not an exercise in post-truth: there is no consensus yet, or the one that exists is still weak. Yet, in the case of nutrition, those who claim that there is a consensus and make very specific recommendations that, like the tide, come and go, state what they do not know as if they know it. 

When a consensus clearly exists, is solid and there is no compelling evidence that puts it at risk, and still part of society does not accept it, we surrender to post-truth. The claim that vaccines are safe and effective is based on evidence from very different fields such as immunology, molecular biology and epidemiology. Regardless, some still reject it. There is also an undisputed consensus that species evolve by natural selection, with very strong evidence coming from many different fields of knowledge. But some people reject this on the basis of alternative ideas, not supported by the evidence. There is enormous consensus around the existence of climate change and the great responsibility that human activity has in it. There is also great consensus that it is safe to consume food that comes from genetically modified organisms. However, there are individuals or groups that oppose these ideas. 

In relation to post-truth, when people start talking about whether or not there is consensus, it is almost always on these issues that are apparently controversial (apparently being the operative word, because in fact the consensus is solid and there are simply people who do not accept it). 

These themes resonate strongly with some people, to the point that they identify them as a constitutive and unquestioned part of the way they look at the world, even as part of their identity. This is another risk when trying to connect with someone who holds positions that are not grounded in evidence, but are weighted in a primary way in the way they conceive the world and themselves: to question these positions is, in a way, to question a part of who they are. This makes the reluctance to change understandable.

When was the last time you tried to identify the ideas that are "part of who you are"? I mean those that you would hardly be willing to change, no matter what evidence is put in front of you. Among other things, writing this made me look for what those ideas are for me.

From outside the world of experts, we can sometimes confuse a controversy about consensus with a side discussion that does not question it. For example, there are minor disputes among experts investigating evolution about whether in some particular case a species evolved in one way or another, whether the mechanism by which something happened was A or B, whether the branching of one species into two different species occurred earlier or later. But these are details that remain to be defined and in no way threaten the enormous consensus that the great mechanism of evolution is natural selection. As part of the fallacies of post-truth, the discussion of details is sometimes interpreted as a crisis of the theory. This is not so: in 1972, Stephen Jay Gould and Niles Eldredge proposed the theory of punctuated equilibrium in evolution and argued strongly with other evolutionists. Even in the midst of an impassioned conversation that involved much of the community in the field, none of the participants were disputing the explanation offered by the theory of evolution based on natural selection; rather, the issue at stake was whether it occurs at a constant rate or whether it speeds up and slows down from time to time. Evolution based on natural selection, far from suffering a crisis in its foundations, was undergoing quite the opposite: a now refined and subtle consolidation, only possible thanks to firm foundations. However, some of those who oppose the consensus on evolution used the discussion to point out that there was no consensus. This is completely false, and is a typical mechanism of post-truth: confusing the fact that there is a conversation with the content of that conversation. 

We will particularly discuss the case of climate change later, but first: why would it matter to us, society, to understand when there is consensus on an issue and when there is not, or how strong a consensus is? Because it is a compass that allows us to answer quickly whether or not something in particular is known or not. This is our first line of defense against the onslaught of post-truth. 

Our knowledge is always tentative, always in development: it is true that we do not know everything, but this does not mean that we know nothing, nor that all narratives about the world are merely equivalent accounts, opinions without an objective correlate in the world. This may cause anguish to those who yearn for the security of absolutely certain knowledge. It is like getting angry with the world because it is not a fairy tale. Being able to separate the wheat from the chaff becomes, in this context, indispensable. As we said before, we need to find the right balance between extremes that do not help us: blind confidence and paralyzing distrust. We cannot be totally sure of anything, but neither should we let our skepticism lead us to think that we cannot act in the world. As David Hume says, "We shall know later whether your skepticism is as absolute and sincere as you say it is, when we finish this meeting: we shall see whether you leave the room by the door or by the window." 4Whether your scepticism is as absolute and sincere as you claim is something we shall learn later on, when we end this little meeting: well see then whether you leave the room through the door or the window. David Hume, Dialogues Concerning Natural Religion.

It is not easy to reach a consensus on a subject, and that, precisely, makes it more valuable. We’re getting to that. 


Generally, within the scientific community (or scientific communities, if we separate them according to the areas they investigate, but unite them by their methods), the aim is to destroy one's own ideas and those of others in a rather aggressive process. This is not necessarily an altruistic act of intellectual honesty: scientists are human, and therefore, they are as competitive, envious, and eager to stand out as anyone else. Who would not want to prove that Einstein was wrong and that one's own theory of relativity is better than his? 

The standard procedure of scientific publication is what we call peer-review: scientists write what they have discovered in a paper, send it to the editor of a journal, who forwards it to other experts in the same field, who act as reviewers and usually remain anonymous. The reviewers can ask questions of the authors, suggest changes, tell them that evidence is missing, that there are errors and, very often, consider that the paper does not have merit for publication. Anyone who has been through the process can attest to its ferocity. And the evidence on which the scientific community relies to reach consensus is this type of publication. In fact, journals have a score that depends on their quality and is related, in turn, to how demanding they are about the relevance of the work and the quality of the evidence shown. Getting scientists to agree is extremely difficult. So when consensus is achieved, it is valuable. Of course, once we reach a consensus, it is not unquestionable, or it would be dogma, not consensus. There may be new evidence, or new interpretations of the evidence, that allow us to come up with a more comprehensive or simpler theory. We do not close our minds to new evidence that could refute that consensus. That is almost the definition of science. But just as we need a great quality and quantity of evidence to create consensus, we also need a great quality and quantity of evidence to abandon it. The requirement, in this sense, is equivalent both to build and to alter a consensus. 

Therefore, even if all our scientific knowledge were to disappear, if every assertion of science were to be contradicted by evidence discovered in the future, the most important thing about science would remain: its methodology based on making hypotheses and then testing them against the real world (Popper’s conjectures and refutations). That, which seems to be a weakness of science, is actually its greatest strength. When a scientific claim is refuted, science has not been refuted. Rather, it has been used in all its power to bring us one step closer to a faithful description of the universe. Scientific theories are refuted using the same methodology. For this reason, if we want to place a marker where the truth lies, always with the broader goal of fighting post-truth, following the scientific consensus helps. 

But following a consensus makes us wonder whether we might not be wrong all the same. The possibility exists, of course. There are two recurring criticisms of the idea of following a consensus. One of them argues that science has already been wrong many times. Let's take two examples: the idea that there were human races superior to others, and that thalidomide, a drug for pregnant women, was harmless to fetuses. 

The idea that there were races superior to others was promoted by some philosophers, writers, opinion makers and also some scientists. This led to concepts such as eugenics and was taken up, for example, by Nazism. This idea never reached scientific status because no convincing evidence was ever obtained to support it. It was just that: an idea promoted by some people. Again, the opinions of scientists that have no evidence are just unfounded opinions, and have as much, or as little value as anyone else’s. 

Thalidomide was different. This drug was marketed in the late 1950s to reduce nausea in pregnant women. Very soon, it was observed that many babies were born with very serious congenital problems. Thalidomide was withdrawn from the market and it was later discovered that a mixture of two variants of the molecule had been marketed: one of them caused harm to fetuses, while the other was effective against nausea and had no harmful effects. A terrible mistake had been made. At that time, it was not yet as well established as it is today that the safety and efficacy of drugs had to be evaluated, particularly those meant for pregnant women and children, nor were there methodological standards in place like today’s. Medicine has advanced a lot in recent decades, and that mistake is very unlikely to happen again, though it is not impossible. Of course, this is not a criticism of the consensus, but of the fact that the evidence, in this case, was not rigorous enough. 

Another criticism that often arises when consensus is followed is that it is generated by hidden powers that stand to benefit from it. Let us look, for example, at the association between cancer and smoking. The evidence that smoking tobacco caused lung cancer took a long time to generate and accumulate. Scientific research is meticulous and precisely because of that it may be slow. Methodologically, the issue was particularly difficult. In the first decades of the 20th century, one of the many strategies of Big Tobacco to market more cigarettes was to try to influence science by funding research that, curiously, showed that smoking did no harm or was even good for one’s health.5 There is so much more about this that it is an entire chapter (X).

Perhaps this is a good time to set limits to the power of evidence. Evidence, as a brick of scientific thought, has its limits, and one of those limits is our values. As a concept, the idea of the existence of races is not based on evidence, but is at best culturally and historically rooted. In fact, the available evidence indicates that human races do not even exist. But suppose evidence were to emerge that certain populations were "better" than others in some sense (and we could discuss what better means). I, for one - and this is very personal - would still believe that all people have the same rights and obligations. I know that, in this example, I would put those values above the evidence without denying it. Perhaps, others would act differently. If we were to put together a basketball team, I would understand if my tallest friend was picked over me. In a way, this implies understanding that evidence is nothing more than information about the world as it is, but not necessarily about the world as we would like it to be.

But once the evidence of the harm caused by smoking grew thanks to a huge amount of independent research coming not only from the "office", but also from epidemiology (lung cancer hardly existed before the 20th century), biochemistry, laboratory animal studies, etc., consensus was generated. 

The consensus was never that smoking was harmless. And it was precisely this consensus, and the discovery that the tobacco companies had known this for decades but had concealed it, which soon made it possible to force these companies to pay more taxes, to warn that smoking is harmful, and to pay millions in compensation to individuals and states (in the United States alone, the payments total some 200 billion dollars over 25 years). If anything protected us from attempts by "harmful corporations" to manipulate the truth, it was that consensus generated in a decentralized and independent manner. If corporations want to influence the consensus, it will not be enough for them to have money and power, because the mechanism that sustains the consensus is not only built on the basis of present evidence and a handful of people, but continues to self-correct with new evidence that anyone can generate, as long as they agree to play the same game and do so with appropriate methods. To think of conspiracies, in these cases, would imply thinking that thousands and thousands of experts in each field are hiding a great secret. But beware: the fact that approaching the truth through rigorous experimentation and discussion with others is almost inevitable does not mean that the attempts of these groups do not hinder or delay consensus building. 

When we see how difficult it is to maintain the slightest secrecy, we understand that the idea of a decentralized and diverse community hiding the truth for a long time is ridiculous. At the same time, when we see how the strategy for managing public opinion has shifted, in some cases, from hiding the truth to covering it up with half-truths, fallacious arguments and outright lies, we understand the real progress and profound danger of post-truth as a strategy to hinder both the creation of consensus and its communication to people outside scientific communities. 

Consensus can be wrong because it is not absolute certainty, but is based on the best evidence available so far. 

"Only a Sith thinks in absolutes": the road to truth is not only steep, but also rocky and winding. A consensus may change. But following the consensus of each era is almost always the best possible decision. Another possibility is to challenge it frankly, and to try to justify that challenge by experimenting, interpreting previous evidence in novel ways, and building a new consensus. But ignoring the consensus because we don't like it and watching the game from the sidelines complaining that it is not the way we want it to be is not proper behavior. 

Now, achieving a certain level of understanding in any specific area takes a few years. To achieve great mastery takes a lifetime of dedication. Citizens do not have the training that experts have (and experts in one area are, in turn, citizens who do not have the training that experts in other areas have). So we may not only not grasp the knowledge fully, but we may also be unable to assess its reliability. But we want to understand, and we want to make informed decisions. How do we do that? Seeking consensus, and following it, helps: the Earth is round, vaccines are effective and safe, making water safe to drink prevents disease, extreme poverty is decreasing in the world.


In this chapter we introduced scientific consensus, the benefits it brings as a beacon for truth and some of its inherent difficulties and limitations. Following a consensus allows us to make decisions based on the best available evidence, even in fields in which we are not experts and therefore lack the ability to judge for ourselves the validity and certainty offered by each particular piece of evidence. 

To some it might seem that following the consensus is a variant of the fallacy of authority, in which we attribute veracity to something based on the authority of the one who holds it, but there is a difference: when we are not experts in the subject, following authorities is inevitable, but these authorities must indeed be such, that is, they must know the facts of their discipline that are relevant to the subject in question and must not have biases or personal interests at stake. Thus, we are actually relying on the evidence through authorities who know and understand it well. In this case, it is not a fallacy, because someone who is not an expert is not invoked as part of the argument. 

But, even if we want to follow the consensus, in real life the statements do not come with a stamped certification, which clarifies whether or not they are aligned with the most consensual view of the subject. So how can we tell whether or not there is a consensus about something, and how solid it is? The experts in each field know how to do this, but how do the rest of us do it? We want to act, and for that we need to know. 

It is time to introduce the third Pocket Survival Guide to guide us, at least part of the way, as we navigate the complexity of consensus with the idea of defining whether or not we trust a statement, and to what extent we do so. New tools are added, then, to our box: a series of questions to help identify consensus, especially on those topics about which we know little or nothing.



Is there independent evidence, quantitatively and qualitatively, to support the claim? 
Was this evidence published in peer-reviewed journals, or other systems that ensure that its quality is similarly endorsed? 
Is this evidence available for interpretation by others? 
 Is the statement aligned with the position of the majority of experts or recognized bodies in the field? 
Are there alternative ideas that have significant support from experts or recognized bodies? 
If there seems to be no consensus, or a very weak consensus, but we need to make a decision, what is the best available evidence?

With this chapter we close this first section, in which the focus was on making several important points in our fight against post-truth: we can know, there are specific mechanisms which allow us to know, and also specific mechanisms to permanently assess and reassess what we already know. Knowing, at least in a very approximate and general way, some of those mechanisms allows us citizens to navigate in the midst of the confusion with a compass that works. This does not mean that we will no longer get lost. The world is complex in itself, and what we know about it is not everything. But we will be better prepared to address the traps that post-truth puts in our path. 

With these tools in our box, we will then move on to the second part of this book, in which we will point out some of the factors that can generate unintentional post-truth, and propose specific tools to recognize them and eliminate them.