09/07/2020

'Science is not full of certainties,' Artur Avila tells G1.

Foto: Divulgação IMPA /Daryan Dornelles

In an interview with Helio Gurovitz's Blog , on G1 , IMPA's extraordinary researcher and winner of the Fields Medal in 2014, Artur Avila, stressed that, unlike political discourse, "science is not full of certainties". In the conversation with the journalist, published this Wednesday (8), the mathematician spoke about the difficulties and opportunities for scientists in Brazil, the ambitions of mathematical models adopted in the pandemic and the origins of anti-scientific thinking.

Artur also addressed the limitations of science during the pandemic, while emphasizing its absolute necessity for understanding the current situation and finding solutions. “The evidence is partial because you simply can't wait for the natural pace of science, which is slow, to make decisions with the level of certainty we all desire. I wish I had more and more data to make the right decision. But inaction can lead to worse consequences. By their very nature, these are not only scientific issues, but societal decisions. All fields, including the humanities, must interact for us to understand how to act in the situation we are in.”

This article is part of the #ScientistAtWork initiative , promoted by the Serrapilheira Institute in celebration of National Science Day, July 8th. Journalists, scientists, and science communicators are being encouraged to occupy highly visible spaces in the press and on social media to talk about the scientific process and show how science works. The Director-General of IMPA, Marcelo Viana, is also participating in the initiative and dedicated his column this Wednesday to the topic.

Check out the full article:

The last time I interviewed Artur Avila, in 2014, he was in Seoul. He had just received the Fields Medal, an honor often compared to the Nobel Prize, awarded to the world's best mathematicians under 40. The only Brazilian to receive this award, Artur is not only one of the country's most successful scientists, but also living proof of how it is possible to train internationally renowned researchers here in Brazil. He earned his doctorate from the Institute of Pure and Applied Mathematics (IMPA) in 2001 and has since divided his time between research in Paris, Rio de Janeiro, Zurich, and wherever else mathematics takes him. As part of the #ScientistAtWork initiative, promoted by the Serrapilheira Institute to celebrate National Science Day, Artur spoke with me yesterday while walking through Parisian streets and boulevards. He described the difficulties and opportunities for scientists in Brazil, criticized the ambitions of mathematical models adopted during the pandemic, analyzed the origins of anti-scientific thinking, and, distracted, realized at a certain point that he was lost. He wanted to go home, near the Place de la Bastille, but the walk had taken him to the Place de la Nation. “I have absolutely no sense of direction,” he said. “I went wherever my nose pointed during the conversation.” But he quickly regained his way. Both in his walk and in the conversation. Below are the main excerpts:

What changed in your life after the Fields Medal? Are you recognized on the street? Do people ask for your autograph?

In an academic career, this puts you in a more relaxed position. I have the advantage of being able to choose where I work and personal conveniences. But, in the end, mathematics is mathematics. Among the sciences, it's the one that places the least importance on hierarchy. There are no arguments from authority. Leadership depends on the ability to make the right decisions every time. A highly recognized researcher can have a dialogue with a beginner student, and if the student is right and shows it, they will have to accept it. This is very good; it's a way of not corrupting the practice. I already had recognition among mathematicians before the medal. What changed was the recognition outside of mathematics, which translates into the fact, for example, of giving this interview, because they may consider me someone to listen to as a scientist. In Brazil, there aren't that many scientists with public recognition. When I'm taken to meetings with young people, then, I have a role that can be motivating, useful for a beginner to have the idea that it's possible for an internationally successful researcher to come from Brazil. On a personal level, when I'm in Brazil, there isn't much difference; I'm not usually recognized on the street.

You are one of the most successful Brazilian scientists. You have also lived abroad for a long time. What are the difficulties of doing science in Brazil? Are there any advantages?

There are specific difficulties in Brazil, and there's a more general, international panorama. We shouldn't have this idea that the life of a scientist outside the country is all that wonderful. It depends on the career level. Abroad, the situation of a young researcher who has completed a doctorate is usually quite precarious. There are no permanent jobs for very long. Many find themselves in difficult situations, going from post-doctoral position to post-doctoral position, with only a few years in between, frequently having to move to another country. It becomes difficult to create family ties. It's psychologically very heavy, but it's the reality. A researcher who has achieved stability will already have a different situation. This is true in Europe, the United States, and elsewhere. In Brazil, there are specific difficulties, which translate into problems and opportunities as well. Science in Brazil is young compared to the major international centers. One difficulty that stems from this, associated with other problems in the country, is that there is no stable research policy. But there is something that can be positive in relation to Europe: this same youth means that there is a much greater potential for growth. In Europe, the situation is saturated, the academic community is consolidated, there is no room to grow. There are only openings for new researchers as others retire. In Brazil, there is potential for growth. It's not being done effectively, but it's something that gives hope.

What makes IMPA, where you graduated, different from the rest of the national scientific community?

It's not a very common institution in the Brazilian landscape. Due to its formation and historical context, it was an institute created with a specific concept by independent researchers who tried to form a small group focused on excellence. It grew with the idea of always reaching a high international level of research. This was possible, in particular, because it didn't need resources to compete immediately. It needed to create opportunities to attract people. But this is less complicated than competing with a laboratory abroad. It's not an institution linked to a university that suffers the same pressures. It's focused on research and postgraduate teaching, although it has since become something that influences mathematics at a greater level in Brazil. Several things have allowed IMPA to establish itself in this position over time. Today it functions efficiently, has international recognition, and practical issues are well resolved. Even if it is affected by what happens at the national level, it remains insulated from the main difficulties that affect other departments and universities.

What kind of difficulty?

Most governments that come to power don't have much of an idea of why science is invested in or how it works. Governments come and go, and you get the impression that the relevant ministry is merely an addendum to distribute positions to political allies, not something central to the country. There's no policy for scientific development. In times of instability, opportunities may arise at a certain point, then the situation changes, and those opportunities are brutally cut short. This affects everyone who had invested their education and their dreams, whose hopes end up betrayed by successive government decisions. My impression is that it's something they think they have to do, but they don't know why or how to do it.

Is there ignorance about the importance of science even among the political elite?

It's also about how science works. They don't even know what it's for or how to do it. The decisions made are incompatible with quality science. You create groups, start projects, then cut them off. All this science that stops halfway through doesn't achieve results. I often use an analogy: the government is sensitive to the idea that it's important to seek stability in the markets, so that investors feel confident and invest in the country. This understanding is easy in economics. Well, the same reasoning applies to science. You won't develop science in a situation of instability. This is true from a personal point of view. Someone with certain characteristics and skills won't invest their youth in one direction when suddenly everything can change completely. You get a doctorate, it ends, and then things no longer exist. It's not reasonable. You can't attract human capital to make a personal investment in that situation. Depending on how it is, people won't agree. The same talent they need to be capable scientists is needed to get more stable jobs. As much as you love doing science, as much as it makes you dream, it might even be madness to pursue a career in such a scenario.

From a work perspective, is mathematics different?

Each scientific activity has its own characteristics. Mathematics is a relatively inexpensive thing to do, compared to sciences where you have experiments. If you don't have the conditions and resources for them, you simply can't advance. In mathematics, you only need the researchers. In most cases, they work with their own minds; they only need a blackboard or something similar. It's a cheaper activity.

Never before has science been so attacked. During the pandemic, denialism has had dramatic consequences. Is there anything we can do? How can we get out of this impasse?

It's not necessary to go into many examples – I immediately think of vaccines – to see that an anti-scientific stance is attractive in certain circles. Why does this happen? First, it's associated with political polarization. People choose their side a priori. They are so attached to their own position that science that challenges it needs to be discarded for them to continue living well. When they refuse to criticize their own ideological field, this leads, in a way, to rejecting science that presents a critique of it. It's part of how people experience politics. Everyone is actually exposed to this. It's necessary to recognize that, on whichever side of the polarization you're on, you will filter information and reject contrary scientific conclusions. It's not a matter of saying that everyone is the same. But, when we criticize – and we should always criticize – positions like flat-Earth theory, it's always good to see if we're not willing to do the same when the criticism comes our way. Political discourse is too shallow to encompass the complexity of a scientific discussion. That's because science isn't full of certainties. Dealing with nuance and how it develops, with margins of error, is complicated. When it becomes a heated rivalry, when you're rooting for the result you want, you don't look at science as a source of knowledge, but only to try to justify actions you've already decided on.

This applies to the pandemic, right?

It's truly a complex issue. The research we need to conduct right now is very difficult. The evidence is partial because you simply can't wait for the natural pace of science, which is slow, to make decisions with the level of certainty we all desire. I wish we had more and more data to make the right decision. But inaction could lead to worse consequences. By their very nature, these are not only scientific questions, but societal decisions. All fields, including the humanities, must interact to understand how to act in the situation we are in. And not just science. Science can only play a part. On the other hand, beyond the nonsense and conspiracy theories, the population begins to see science as a resource in such a situation. Scientists are the ones who will help. While not miracle workers, they are our best hope. This reinforces the importance of science.

It is also a challenge for the scientists themselves, who are not used to doing science in real time…

It's not what I'm used to, at least (laughs). I do pure mathematics, in a time when there are no deadlines. My problems don't respond to the pressures of the real world. But, like many people who aren't epidemiologists or specialists, I became interested in the topics and tried to understand what the difficulties were. First, there are methodological problems: how to obtain quality data while people are dying? It's not the ideal situation to get very clean data, and you have to deal with what you have. This can also lead to theoretical questions: how to deal with this bad data to reach some worthwhile conclusion?

How does your research area, dynamical systems, relate to what we are experiencing in the pandemic?

I like to joke that when I look at gravitational equations, I formulate questions that relate to timescales much larger than the lifespan of the Solar System. These are models where the fact that the Sun exploded doesn't interfere with what I see. Obviously, looking at the pandemic, interested in estimating what the real situation was, I tried to consider the possible difficulties. In my analytical capacity, I observed that, in general, there is a tendency for modeling to be done excessively. The problem is sometimes having too much confidence in the models. There are extremely complex models, too complicated for the current situation, because the uncertainties overwhelm their capabilities. I say this guided by dynamical systems. When I model what happens with extreme precision, I am perfectly aware that this very fine analysis will only be valid if this model holds true. If there is a small variation, perhaps a more robust analysis is needed. People think that if they put more and more complexity into the model, they can have more confidence. This is not necessarily true. You need to accept uncertainty. Think about the problem we are trying to understand: the progression of the pandemic in its initial stages. It would be crucial to be able to provide a valid forecast for a few weeks from now, to guide resource allocation. That would be desirable. On the other hand, the nature of the problem introduces such a high level of uncertainty that it renders the resulting numbers somewhat useless. The nature of human interactions and the heterogeneity of relationship networks may be far more relevant. People were trying to find numbers that represented transmission, but what might be most important are small details of the interaction network, especially at a time when the virus is less prevalent. Models become much more robust when a large part of the population is already infected; then it becomes more reasonable to assume certain uniformities in distribution. We want to provide the best possible information, but what science often teaches, when I look at real-world problems, is that you should be concerned with the limits of what you could know. You should try to identify the point beyond which you will no longer be certain, no matter how hard you try, no matter how much you try to model. I've reached that point. It's something understood in weather forecasting. Uncertainty means that, no matter how many new weather stations you add, there comes a point – and it's not far off – when you can't tell if it's sunny or rainy. The loss of information is very rapid; you can't fight against exponential effects. The pandemic situation isn't much different. But the knowledge that science provides, even if it's negative knowledge, is useful because it lets you know how to react. We know we'll have to make a decision in the face of an inherently uncertain situation. Maybe it's not the information we'd like, but it's something.

Those who look to science expecting truth, as a kind of oracle that separates right from wrong, are fundamentally deluded. How much of denialism doesn't originate from this absolutist expectation of truth? After all, science doesn't have answers for everything.

We have a lot of uncertainty, especially in this situation where we're doing science in real time. It's not the natural pace of science. Personally, I reserve the word "denialism" for well-established situations. There's so much uncertainty at the moment that using it too early ends up associating it with clearer situations, where scientific consensus is much greater. It's not that I doubt science, I don't have any. But it encourages people to react more virulently. Science, especially at this moment, will appear to oscillate. One moment scientists will recommend one thing, another moment they may recommend the opposite. It's not good to prematurely use this terminology in a situation where things can change. The problem with those who adopt an inherently anti-scientific position is that, by complete accident, they might suddenly be right by chance. If the recommendations become the opposite of what they were, it doesn't mean that science was wrong. Science was simply expanding knowledge. Using this terminology, however, will encourage people to interpret science as actually having a political function. That's why I try to avoid it. It's vocabulary that can encourage anti-scientific thinking. We have to accept uncertainty and say that there's a range of possible conclusions based on our current knowledge. And we should reserve denialism for clearer situations, like the Holocaust, which is very different from issues that scientists are still debating. We need to avoid excessive use of language because we think we're helping to clarify something, but it can have the opposite effect: leading people to doubt even more, because the evident uncertainties can give the impression that science isn't a serious matter.

What would be more important for laypeople to understand about how science or mathematics works?

We use the term "science" for many different things. My particular area, which I can speak about with authority, deals with totally abstract concepts, with little concern for possible applications. The work of the mathematician, in this almost artistic conception, is useful in the development of other, more applied fields of mathematics. Discoveries in certain particular contexts, dissociated from applications, can later be essential in applications. This happens all the time. It also works more generally. Diverse scientific areas can benefit each other. Physics benefits from mathematics. Mathematics benefits from physics. Physics creates problems that will attract the attention of mathematicians and open perspectives that would not have been discovered in isolation. There is this coherence. When thinking about the development of science, it is particularly important to see it as a whole. It is wrong to try to chart too early in which direction research should go. It may be tempting, if we have limited resources, to direct those meager resources to try to do useful things in a more direct way. But this will not lead to the best results, even within this limited perspective. Simply because you don't know. Original research, the kind that can truly be called research, will only answer what you didn't know before. It involves a process of discovery where you don't know where things will come from. They may come from another area that you could never intuit. Therefore, there needs to be broad development to answer even practical questions, even those that seem quite clear. I'm generally speaking in relation to mathematics. You need to think about unified mathematics; you can't think about pure and applied mathematics separately. Dialogue is necessary because it's dialogue that moves things forward. The same perspective applies more broadly to the interactions of mathematics with other sciences and among the other sciences themselves. This is something that needs to be understood when making any science policy. People think that science exists to answer precise questions, that it's possible to provide resources to answer only those questions. But then you don't answer them, or anything else, because you're not doing real science.