It’s not what you know, it’s who you…. trust

Think of something basic that you know.
 
‘The earth goes around the sun’, for example.
Or the ‘Central Dogma’ that genetic information flows from DNA to RNA to protein.
 
How would you demonstrate these from first principles? Apart from a few astronomers and nucleic acid experts, the only honest answer for most of us is ‘we can’t’. And very few indeed could do this for both.
 
So why do we think we ‘know’ them?
 
Pause for a moment before you read on. Can you remember exactly how you learned each of these and how it felt at the time? And what makes you continue to believe them? This might seem abstract, almost philosophical, but it is fundamental to understanding what we mean when we say we ‘know’ something in science. And understanding this can reduce the anguish we feel when events contradict our existing knowledge.

Trust is everything
I don’t remember the specific moment when I learned that the earth goes around the sun, but I do remember a phase of general fascination with the solar system, sparked by watching the Apollo moon landings, and rushing off to the local library to devour book after book on the topic. It’s likely I learned it somewhere during that period.  

​For the ‘Central Dogma’, in contrast, I have a very vivid and specific memory of the moment I learned this. Four decades later I can still see and hear my secondary school biology teacher telling us as if it was yesterday. It seemed important, not just to pass my exams but to understand the fundamentals of how our cells work, so it clicked into place in my memory right away.
 
The point here is that we don’t know in the true sense of the word. We trust. I trusted the library books, especially when they all said broadly the same thing, and I trusted my parents and teachers who backed that up. Years later doing my ‘A’-levels, I even understood the gravitational forces to explain planetary orbits. Yet I have never demonstrated this experimentally, nor would I even know where to start doing so. I just trusted the people and the books I learned from and I’ve had no serious reason to question it ever since.
 
And similarly for the flow of genetic information. Although I learned some of the supporting evidence, I have never done these experiments myself. I’ve seen the concept challenged a couple of times, learning about reverse transcriptase as an undergraduate, and about prions as a PhD student, but apart from these exceptions, all my subsequent understanding of molecular and cell biology fits the ‘Central Dogma’. The research in our own laboratory today, and in many others, would be impossible if it were wrong. But I have never tested it directly. I just trusted the teacher and never questioned it further. It explains what I see and acts as a good basis for experiments.

Alternative facts
Here’s the problem with this approach. Flat earthers, creationists and anti-vaxxers also trust the person who told them these ‘alternative facts’! So what makes us different? Like us, they constantly have experiences that support their beliefs - others who say the same thing or people who react badly to a COVID vaccination, for example. They explain away anything that conflicts with their belief (science, in other words!) just as we explain away their beliefs saying “they are crackpots”. But are they? Or are they doing the same thing we do – trusting what they have been told, reinforcing it when they see something consistent with it but without the ability to directly confirm it?
 
Could it just be that they trust different people because of an accident of birth? After all, even someone brought up on a heavy diet of conspiracy theories can change their views if they are given the social safety net to do so.
 
By now you may be feeling disorientated. But that’s the point – to be on solid ground we need to think hard about who, or what, it is that we can genuinely trust.  

What we trust
Ultimately of course, it is not people we trust when we say we ‘know’ something, it is the self-correcting process of science itself. Overturning received wisdom brings great personal satisfaction and career reward so  someone somewhere will be trying to do this. Models that withstand this process for long enough are almost certainly correct, although they may still be oversimplifications of reality. But how much time has to pass before we can consider them 'facts'? The same biology teacher who taught us DNA to RNA to protein also taught us ‘one gene, one protein’. Alternate splicing had been discovered a few years earlier. I don't know whether he knew this already but it was clearly not yet on the syllabus. Indeed it was many years before anyone realised quite how widespread it is.   
 
How rigorous is this self-correction process and how long does it take? It can be extraordinarily hard, and take a long time, to overturn received wisdom. And personality still plays a role. Someone perceived as an ‘outsider’ is, sadly, less likely to be listened to – one of many justifications for us to address what is commonly called ‘unconscious bias’ (the topic of a forthcoming article). Barry Marshall famously had to resort to drinking H. pylori bacteria to overcome the decades old doctrine that excess acid causes stomach ulcers, work that led to his Nobel Prize. Not every ground-breaking research project gets the recognition it deserves in papers or grants. How many quit science before they get the recognition they deserve? Even peer review is imperfect so fundamentally flawed papers sometimes get into so-called ‘top’ journals, perhaps with even greater incentive to cross ethical boundaries for such ‘high profile’ publications. Using a journal’s reputation as a proxy for quality isn’t as reliable as it may seem. And just as flat earthers selectively interpret science, which of us, in all honesty, is as pleased to see data that disprove our hypothesis as data supporting it, especially when the latter is more likely to advance our career?

​So while it is true that science is a self-correcting process over the long-term, ‘long’ may mean periods longer than our career - not much use if you're hoping what you see as a breakthrough will help you get tenure. So how else can we know what to believe?

Trust in people
Until this bumpy process runs its course, how can we know where to place our trust? Trusting data that fit best with what we already believe doesn’t work because the whole point of research is to test whether existing models can be overturned. Basing trust on this really would make us no different from flat earthers. We can weigh the strength of the data supporting a new result against the data supporting our existing model, and ideally try to replicate some of it. However, this is limited to fields we know well enough to do this and we are unlikely to have time to replicate everything. Another marker is when papers are published from independent researchers that support it, but even here we need to be wary of the ‘bandwagon effect’, something I have definitely witnessed.
 
What about the person, or research group, involved? Could this be the basis for trust? Over many years I’ve become increasingly wary of focussing on ‘big names’ in a given field. Could they, for example, have become ‘big names’ through dominant personalities, having the right contacts or, worst of all, fraudulent papers that have not yet been exposed? What impresses me far more is groups that have consistently published work that other people can build on.
 
We need instead to go back to the fundamentals of the markers of trust: body language, openness, humility, competence, a willingness to listen to other views. All of these are best assessed in person, one of the main reasons why in person conferences are indispensable. Nothing informs us of a scientist’s integrity quite like seeing whether a conference speaker answers questions directly and openly, or dodges them like a slippery politician, and seeing their body language as they do so.

The modern world
Crucially, video conferencing is a poor substitute for these, especially with people we have not met previously in person. Information from the Internet, social media or artificial intelligence tells us even less about any of the above. We need to be particularly careful that our emotional reaction to the unquestionable speed and power of AI does not trigger us into believing its output too readily. My guess is that, in time, we will slowly learn to trust some AI providers more than others but we are barely at the beginning of that journey right now with all kinds of risks of picking up false beliefs when there is apparent authority but no body language to guide us. 

Reducing anguish by understanding the trust-knowledge link
What we call ‘knowledge’ then is mostly a proxy for trust. This may explain why having it challenged can be so upsetting: to question our knowledge is to question our trust in the people who told us. But if we recognise this link and the need for a Bayesian approach to knowledge, constantly updating our beliefs as new information comes along this becomes easier to take. Instead of looking for false certainty, we can begin to see all knowledge as greyscale, with degrees of confidence that can be updated and occasionally radically overturned.
 
We need to feel the freedom to be wrong. Not only is this the key to new learning. It can give us a more peaceful life too.