I love reading about the history of science. In part, because I think you can learn so much about the process of thinking itself. Especially when it goes wrong. More especially when you are looking at the process of immunisation.
Immunisation is something that Karl Popper was particularly interested in. Popper was a scientific philosopher who is a bit of a hero of mine (when I can actually understand what he is saying). Amongst many other things, he was interested in the techniques used by scientists to protect favoured scientific hypotheses, which he called ‘immunisations’.
An immunisation is essentially a way of explaining why a fact, which appears to contradict a favoured hypothesis, does not actually contradict it at all. For example, when it was found that the orbit of the planet mercury could not be explained by classical Newtonian physics, a mathematician called Le Verrier postulated that there must be another, smaller, planet inside the orbit of Mercury that was affecting Mercury’s orbit. The planet Vulcan.
Vulcan was invisible – primarily because it did not exist. But for many years the invisible and non-existent planet served its purpose. It protected classical Newtonian physics from a potential contradiction, or refutation. Or, to be more blunt, of being simply wrong. In this case, scientists were quite happy to believe in invisible non-existent things, if the alternative was to cast aside a hallowed hypothesis.
Of course, this is just one of thousands of examples whereby unwelcome facts have been simply swatted aside, or immunised against. It is not just the Catholic Church that refuses to look through telescopes.
Vulcan, although just one example, does provide a good case study of a widely used form of immunisation tactic, the ‘ad-hoc’ hypothesis. An ad-hoc hypothesis is a secondary hypothesis that is bolted on to the side of the main hypothesis in order to defend it, or protect it. A more recent example of this can be seen in the Global warming debate.
It has been noted that global temperatures have not increased by much, if at all, in the last 15 years. This, however, is not viewed as a contradiction to the hypothesis of man-made global warming. Why not? Because it is argued that the oceans are taking in the excess heat, and trapping it. This process has held back the degree of global warming that had been predicted by the experts.
I am not going to debate whether or not this is true. I am just using it as a more recent example of an ‘ad-hoc’ hypothesis which came into existence to protect the central hypothesis. I would further add that ad-hoc hypothesis are not always wrong. They can very often be right. Le Verrier, prior to inventing the planet Vulcan, had predicted the presence of the plant Neptune due to irregularities in the orbit of Uranus.
However, if you read his works, you will know that Popper was not a fan of ad-hoc hypotheses. He felt that a good hypothesis should be fully predictive of future ‘events’ without the need for additional explanations, adaptations, or suchlike.
He did not state how many ad-hoc hypotheses it took before you had to admit defeat. One, ten, a hundred, a thousand? No-one can give you a clear cut figure, but the more of them there are, the less likely it is your central hypothesis was correct in the first place. The phenomenon of adaptation/immunisation had been recognised many years before Popper.
‘A nice adaptation of conditions will make almost any hypothesis agree with the phenomenon. This will please the imagination, but does not advance our knowledge.’ J Black 1803
I have recently been pondering the ad-hoc hypothesis once more in relation to heart disease. For I suspect that never in the history of science has a central hypothesis had so many ad-hoc hypotheses bolted on to it. Indeed, we have now reached the point where ad-hoc hypotheses have had ad-hoc hypotheses bolted onto them, to protect the ad-hoc hypotheses themselves from being refuted.
Just to look at one example. There are a number of drugs that have been developed to raise High Density Lipoproteins (HDL), the supposed ‘good’ cholesterol. A few of them also lower LDL ‘bad’ cholesterol at the same time. Billions have been spend on this class of drugs known as ‘trapibs’ . The first of these was Torcetrapib.
At this point I should probably remind you that the ‘good’ cholesterol hypothesis was only created as an ad-hoc hypothesis to explain why some/many people with high total cholesterol levels do not suffer from heart disease. ‘It’s because they have a high HDL level.’
The logic here was obvious, if horribly facile. Raise the HDL and reduce the risk of heart disease. Anyway, ignore the chasms of logic, along came the ‘trapibs’, which were going to take over from statins:
‘Hailed as a potential blockbuster that could take Lipitor’s place, torcetrapib was a cholesteryl-ester transfer protein inhibitor (CEP-T inhibitor) designed to increase good cholesterol and lower bad cholesterol. Development of the drug began in 1990 with clinical trials starting nine years later. But it wasn’t until 2006 that Pfizer got close to submitting the drug to the FDA. The company touted torcetrapib as the answer to its near-term pipeline woes, predicting the potential blockbuster could make up for billions of dollars in lost Lipitor sales when that drug went off patent in 2011.’
Well, torcetrapib certain raised HDL by about 50%, and lowered LDL by about 10%. So, what could possibly go wrong?
‘What went wrong: In late 2006, the walls came crashing down around the company. Pfizer announced in December that it was halting development of it’s prized Phase III asset. The decision came after an independent Data Safety Monitoring Board recommended terminating the study because of an imbalance of mortality and cardiovascular events. Results from a 15,000-person trial showed that patients taking torcetrapib with Lipitor experienced excess deaths than those taking Lipitor alone. Not long after torcetrapib demise, Pfizer announced that it was cutting 10,000 jobs. The company spent $800 million developing the drug.’ http://www.fiercepharma.com/special-reports/pharmas-biggest-flops/torcetrapib-pharmas-biggest-flops
What went wrong was the Torcetrapib increased cardiovascular deaths about around 50% (relative increase in risk). Several other ‘trapibs’ have since come, failed, and slunk from the playing field, taking many billions down the drain with them. Yes, I know, you have never heard of them. At the risk of sounding rather big-headed, I predicted their total and abject failure long before the results of the clinical trials came out.
Now, there are those of us i.e. me who would suggest that this blows a hole in the entire good, bad, cholesterol hypothesis. But no. Why not? Because it was found that torcetrapib raised the blood pressure, and lowered potassium levels. This, it seems, was enough to explain the massive rise in CV mortality. Well, quite reasonable, you might say. Yes, but the rise in BP was minute, and the drop in potassium was equally minute. This could explain, perhaps, a 5% rise (at most) in CV mortality. Which should have been overwhelmed by the massive rise in HDL, and drop in LDL.
But no-one was going to look too closely into the figures themselves. An ad-hoc hypothesis had been found. The ‘experts’, rather than questioning the central good/bad cholesterol hypothesis simply bolted on the ‘BP rise, potassium falls’ ad-hoc immunisation device and moved on.
So, here we have an ad-hoc hypothesis, bolted onto an ad-hoc hypothesis, bolted onto the central hypothesis. We have another planet inside the invisible planet Vulcan, to explain why it is so difficult to find the planet Vulcan.
As you can see, the games played to protect the cholesterol hypothesis are, literally, endless. I am not sure when the games end? Perhaps they never do. Very clever people, given enough time and money can, it seems, twist reality round and round, inside out and upside down forever. I would call the process vulcanisation, but I think that has something to do with rubber.