Saturday, 28 May 2022

Four Reasons Not To 'Trust The Experts'

A standard reaction to the disastrous democratic discourse and political mismanagement of public interest issues from the economy to public health to gun control has been to demand more respect for experts. I am sympathetic to the idea that when it comes to facts it is better to look them up than to try to work them out for ourselves, and that the way to do that is ask the experts: people in good standing in the relevant epistemic community (previously: Democracy is Not a Truth Machine). 

Nevertheless, there are problems with the 'trust the experts' mantra that should be acknowledged if we are not to fall into an epistemic trap of misplaced faith. Here are four that I try to keep in mind.

1. Too Single-minded
The most obvious problem with experts is the very thing that makes them useful to us. They know an abnormal amount about a very specific issue because they spend a very great proportion of their waking hours investigating it and discussing it with other investigators. Since so much of their own world revolves around that issue, it is not surprising that they tend to think the world does too. The problem is that their very expertise has made them less able to step back and see the real practical significance of what they are expert in. Think for example of the dermatologists declaring that we should wear sunblock even on cloudy days so as to reduce the risk of cancer, or the epidemiologists recommending extreme self-isolation and school closures during Covid.

This inability to see the forest for the trees explains why experts recommend excessive spending on the problems in their area. They are misled by their disproportionate knowledge of certain risks to underestimate the relative significance of other issues, and therefore recommend policies that are a poor use of limited resources. It also explains why many experts are rather bad at making predictions in the many cases where outcomes are interdependent (i.e. not a closed system, like chess or physics). Their expertise on specific causal factors actually makes them less able to appreciate the significance of other very relevant factors. 

The upshot is that bona fide experts are not oracles! They aren't capable of telling us what to do on any complicated issue whether gun violence or Covid or climate change. They are better thought of as resources that need to be managed effectively by their users to be of value.

2. Experts on orthodoxy 

So far I assumed that experts know something relevant, but have problems with proportionality. But it is also possible for experts to be partly or completely mistaken about what they know. Specifically, the nature of expertise is that you are taken seriously by other relevant experts on an issue. But this only demonstrates a shared understanding of the methods and findings of the epistemic community. It does not demonstrate that the epistemic community itself is approaching the issue in the right way. Unfortunately, the history of ideas is littered with epistemic communities manufacturing what turned out to be worthless expertise, such as astrology, theology, Marxian economics, psychoanalysis, Aristotelian biology, or Ayurverdic medicine. 

The problem is that all experts are appointed by and to orthodoxy. Consulting an expert only gives you a window on the current orthodoxy, and consulting more experts only gives you more windows on the same orthodoxy. Whether or not the current orthodoxy is reliable is an independent question on which the experts cannot give you guidance. 

Fortunately, the scientific method has revolutionised the quality of certain epistemic communities by inculcating an obsession with competitive rigorous empirical argument (Michael Strevens). This permits well-functioning scientific communities to continuously revise and improve their methodologies, so that we can have some faith that the inevitable errors (such as sugar vs fat) will be found and corrected. We can also have reasonable faith that we can't do any better than the experts. The relevant challenges will already be being discussed and debated within the epistemic community and experts will already incorporate the prevailing judgement of their significance in their pronouncements. Outsiders like us cannot do better than to trust the experts, since simply knowing of a challenge does not allow us to evaluate its credibility or importance.  

However, a further problem appears. There are often various orthodoxies to choose from and it is therefore unclear which kind of expert to consult. For example, which expert should you consult about whether there are such things as miracles? A philosopher? A physicist? A sociologist? A psychologist? A bishop? A theologian? All these can reasonably described as experts from a relevant epistemic community, but that is not enough to determine whether their expertise is worth having.

3. The Circularity of Self-Selection 

A further problem that relates to both the preceding is that although experts are appointed by and to orthodoxy, they themselves select which orthodoxy to pursue based on their prior interests, values and beliefs. Hence theology attracts people who believe that there is a God and that studying and discussing Him and His sayings is very important - more important than other things they could do with their lives. The resulting epistemic community will unsurprisingly be overwhelmingly made up of quite religious people - albeit of a rather intellectual sort - and this will shape the character of that community, such as the questions it finds interesting and the kind of answers it takes seriously. 

In the same way, gender studies attracts people who identify as feminists and see gender as a social construction - and who tend to disagree with the way gender is studied in other social sciences. Decolonisation studies attracts people who think that white supremacy is still a thing - and not those who think that is a mistaken approach. Psychologists and doctors who specialise in transgender treatment are overwhelmingly 'pro-Trans' - because why would you get into that field if you didn't believe in these treatments and their importance? 

The problem that comes out of this is that expertise is vulnerable to circularity. All you need to become an expert is to be part of an epistemic community that takes you seriously, but all you need to establish an epistemic community is for enough people who see the world the same way to get together and found a journal. When we see a survey showing extraordinary consensus among 'the experts' on some point, we should bear in mind that this includes all the people who were willing to adopt this orthodoxy, and excludes all those who were not. 

If we have a well-functioning (i.e. scientific) epistemic community then we can still have faith that that consensus means something beyond signifying criteria for membership of the group. But many epistemic communities do not strive to challenge and improve their assumptions and methods. This is why you often find them institutionally located outside traditional scientific disciplines, safe from the scrutiny of non-believers. At its worst, self-selection can render expert consensus entirely worthless, resembling the time when Prime Minister Modi surveyed the people who had downloaded his personal app on whether they approved of a controversial policy strongly associated with him.

4. Experts Are Not Natural Communicators

By definition, experts have an abnormal interest in an issue that causes them to spend a disproportionate amount of their time investigating it and talking to other investigators. Their expertise therefore comes at the expense of those other things they could be doing but find less important. One of the other things that experts could be doing with their time instead is explaining the issue to non-experts. This is necessary if anyone else is to benefit from their expertise, but from the expert's perspective it comes at the price of taking time away from their pursuit of what they are most interested in. So the question is why is a real expert wasting their time talking to us and what quality of communication should we expect?

The first problem is that experts pay a price (opportunity cost) for sharing their expertise with non-experts and that price rises the higher their standing in the epistemic community. People at the top of their fields are making contributions that their peers find interesting and important and have important roles in managing the community (as journal editors, department chairs, etc). Taking time away from this comes at a higher price than if they were less successful i.e. less expert in the field. It follows that, other things being equal, the people most willing to share their expertise outside the field will be those with the least expertise to share. (For example, you get to hear from me rather than the highest performing academic philosophers - who are too busy jetting between prestigious conferences to give the keynotes everyone came to see and writing the books that everyone has been waiting for to write blogs.) 

A second problem is that the sharing of expertise is itself a complex skill requiring talents, interests and practise - which we should not presume to expect from experts. 

First, successful communication requires particular talents (such as the empathy and patience it takes to explain complicated things to idiots and answer their stupid questions, over and over again) different from those required for being a successful and respected investigator. 

Second, it requires that the communicators have some interest that would make the exercise worth their personal efforts and costs. But many of those motives undermine the trustworthiness of the expertise that is being shared. For example, some experts receive large amounts of money  - which has been shown over and over to bias their judgements to suit their paymasters. For another example, 'activist' experts want to turn their knowledge into political action, at the risk that their motivated reasoning will undermine their epistemic judgement (previously). For yet another example, many public spirited experts are motivated by an interest in persuading people to do particular things, rather than merely to believe particular things (i.e. practical rather than theoretical reason). Think about how public health experts tried to get people to adopt particular behaviours during the Covid emergency. But also think about how that went wrong. Experts on epidemiology turned out not to be experts in applied social psychology. They also - rightly - lost credibility when the rationalisations they had concocted to make people do certain things unraveled.

Lastly, successful communication requires deliberative practise, the opportunity cost of which is particularly expensive for real experts because it is their high productivity within the epistemic community that generates their standing. Hence we should expect a roughly inverse relationship (other things being equal) between the quality of an expert's understanding of an issue and their ability to explain it clearly to the rest of us.


To sum up. Experts know more than us about important things we would like to know about (though not enough to study it for ourselves). They are an essential resource. We should not indulge the fantasy that we can think well about these issues without their help, let alone that we know better. Nevertheless, it is also a fantasy to think we can or should outsource our thinking on these issues to experts. There is no escape from the responsibility to think properly for ourselves.

Update: Nice tweet on this