Healthy hypothesis attachment

I recently came across the pamphlet-style text ‘Advice to a young scientist’, by Peter Medawar (1915-1987). Medawar was awarded a Nobel in 1960 for his work on acquired immunological tolerance, deeply related to the development of organ and tissue transplants. I came across the book while reading ‘Why we die’ by Venki Ramakrishnan, a fun book I highly recommend to anybody as ignorant as myself about molecular biology. Medawar appears to be an astonishingly entertaining writer, and I’m sure his messages are as poignant now as they would have been in 1979. ‘Advice to a young scientist’ is an omnibus of advice and observations, taken from his own experience and readings, about life as a scientist.

I have read about half of it so far. In addition to a few interesting tidbits that I had never come across before - William Whewell having coined the words ‘Scientist’, ‘Physicist’, ‘Anode/Cathode’, and ‘Linguistics’, and the phrase Scientmanship was wholly new to me - the section ‘Mistakes’ incited the longest pause. I highlight one paragraph:

“I cannot give any scientist of any age better advice than this: the intensity of the conviction that a hypothesis is true has no bearing on whether it is true or not. The importance of the strength of our conviction is only to provide a proportionately strong incentive to find out if the hypothesis will stand up to critical evaluation.”

Seemingly obvious advice of course, appearing in many guises in many places. Yet I think it warrants a bit of attention on my part after witnessing a number of practical failures in application, and given how it amounts to a kind of core scientific principle that I aim to internalise.

The dangers of forming too strong an attachment to a hypothesis are clear. The job of a scientist is surely to carve out a portion of truth from a confusion of observations and experiments. It so happens that one very useful mode of truth-seeking is the hypothesis, which we can offer and then shoot down in turn. Our hypotheses ought to be nothing more than a link in the chain to the prize that is true knowledge and understanding.

And yet we grow attached to them and find it hard to let them go, sometimes slowing down our progress by our inability to acknowledge the falsity of our own cherished hypothesis. Medawar acknowledges this attachment as being instrumental in its own right. It is true that our fondness for a hypotheses does provide a natural incentive to find out if it will stand up to brutal interrogation. In and of itself this is only positive. I suppose danger comes about when we confuse an incentive to find out if the hypothesis will stand up to critical evaluation with an incentive to see the hypothesis stand up to critical evaluation. Simply, the goal must be to interrogate our hypothesis within and inch of its life and see if it still stands, nothing more and nothing less. We must be careful not to interrogate our hypothesis with the intention of demonstrating it’s resilience to interrogation - this is sure to fail.

It is undoubtedly a challenge, and it’s a battle that we must fight within ourselves at first measure. I think it’s very easy to acknowledge this state of affairs, say that you will be careful, and then to go ahead and completely fail to scrutinise cherished hypothesise to the required extent. I can think of many examples of this in my own research. Nonetheless, scientists do on bulk succeed at this. I think that this kind of intellectual honesty with oneself is a very endearing quality of scientists. Something I try consciously to exercise is an interrogation mindset - where I play devils advocate with my own hypotheses and try to get pleasure from the knowledge that I am actively seeking the truth of the matter. I find this becomes more and more of an ingrained habit with time, so I imagine at some point it might even stop being a conscious act altogether - this is the impression I get from more senior researchers.

Speaking of colleagues, discussions and presentations are clearly the powerhouse for avoiding unhealthy hypothesis attachment. Having your ideas interrogated by collaborators, or even better random colleagues who aren’t invested, is a good way to gauge hypothesis attachment; if you find yourself unable to convince others of the strength of your hypothesis, then perhaps you shouldn’t be so convinced either. Thankfully, that this happens is the default unless you work in total isolation. I suppose the only way you can mess it up is by being too stubborn in the face of alternative views.