Isaac Asimov once said: “in Man is a three-pound brain which, as far as we know, is the most complex and orderly arrangement of matter in the universe”. It’s rather ironic that the most complicated thing we’ve ever discovered is the very thing that does the discovering.
To anyone who thinks that they’ve got the brain all figured out, I offer the placebo effect. Find someone who is in immense pain and inject them with the principle active ingredient in opium: morphine. Its proven pain-relieving properties will soon have things under control. After several days, and without telling the patient, replace the morphine with a saline solution – i.e. instead of:
(5-alpha,6-alpha)-7,8-didehydro-4,5-epoxy-17-methylmorphinan-3,6-diol
(a.k.a. morphine), give them table salt and water. The patient, astonishingly, will report that their pain has been relieved.
Things get weirder. Fabrizio Benedetti of the University of Turin in Italy repeated the above experiment, but combined the saline solution with a drug that blocks the effects of morphine. Suddenly, the saline solution loses its pain-relieving powers. This suggests that the placebo effect has at least some biochemical factors.
Curiouser and curiouser. Take the following experiment, as reported by Michael Brooks in New Scientist (23/08/2008). You have a drug (called CCK-antagonist) whose pain-relieving properties you want to assess. So, like a good medical researcher, you perform the classic “double-blind-randomised-controlled trial”. Randomly chose half of the test group to get the drug, and give the other half sugar pills, but don’t even let the researchers know which group is which until after the results have been processed. In this case, CCK-antagonist performs significantly better than the placebo. Great!
Now repeat the trial, but don’t tell the subjects that they’re receiving a painkiller. Suddenly, CCK-antagonist has no effect! It only works if they know it’s a painkiller. Benedetti suggests that it is “placebo amplifier” – if your body releases endorphins (happy hormones), the CCK-antagonist will stimulate it to produce still more.
The weirdest thing about the placebo effect is that it’s the sort of thing that seems anathema to science, almost something out of fairy tales. If I told you that I can see fairies, but that you can only see them if you believe in them, you’d think that I’d forgotten to take my medicine. And if you told me that a certain medicine only works if you believe in it, I’d think you were off with the fairies!
The cynic in me quickly reaches the following conclusion: we don’t have a clue what these drugs are really doing, and we don’t have a clue how the brain really works. We’re just pumping enough people with enough chemical cocktails that we eventually get lucky. This is not the sort of science that I am used to. In the physical sciences, we employ a model of science known as hypothetico-deductivism: we postulate a hypothetical model, deduce the predictions of that model and then see if nature confirms our predictions. For example, we postulate that the universe is homogeneous and isotropic (i.e. is the same everywhere and looks the same in all directions), we deduce that the velocity of other galaxies would then be proportional to their distance, and then confirm the prediction with redshift observations. If medical science worked this way, we would have a theoretical model of a human being in pain, to which we would add a theoretical model of CCK-antagonist, and we would make predictions of its effects.
But these drug trials seem to suggest that medical scientists use a purely inductive model – just make all the observations you can and try to infer a pattern. This was the concept of science promoted by Francis Bacon in the 1600’s. Bacon died of pneumonia, which he caught while stuffing a chicken with snow to test its meat preserving properties. Such a scattered approach to scientific investigation is typical of Bacon – his notebooks contain painstaking observations of the change in the temperature of a dung heap with time. The problems with this inductive approach are neatly illustrated by the placebo effect – it’s very hard to discern cause from effect, to generalise from specific cases, to infer mathematical equations from their solutions, to predict what will happen for the next dung hill or the next drug. The laws of physics that govern heat are partial differential equations; they are difficult to derive using only a thermometer and a cow pat.
We arrive back at the quote I started with. Oddly enough, the universe as a whole is much, much simpler than the brain we are using to understand it. A truly predictive, theoretical model of a human being (at least on a biochemical level) would be a wonderful thing – we could predict the effects of a drug before we even synthesise it. But life remains, it seems, more difficult to predict than the whole universe. Martin Rees said it well: “for all we know the earth, tiny though it is, could be the centre of the cosmos in terms of complexity.”
Surely those two models of science are not as mutually exclusive as you make them out to be – one cannot construct a hypothesis or postulate a model out of thin air, one must start with a series of observations, or else you have nothing to go on. We don’t just postulate that the universe is homogeneous and isotropic without seeing that our bit of the universe gets more and more homogeneous and isotropic as we look at larger and larger scales.
Unless of course I completely misunderstand some fundamental point, which is not unheard of!
I think you’re right – we must have some idea about the world before we can postulate a model. But it’s interesting that, historically, cosmologists did simply postulate that the universe is homogeneous and isotropic (the cosmological principle) before they knew that the universe gets more and more homogeneous at larger scales, and before the staggering isotropy of the CMB was discovered. The cosmological principle was postulated purely for its simplicity. Indeed, this sparked major debates in the 1930’s and 40’s: inductivists like Herbert Dingle opposed this approach, calling on his fellow cosmologists to “[call] a spade a spade and not a perfect agricultural principle”. Read more here: http://www.seop.leeds.ac.uk/entries/cosmology-30s/
I suppose we can think of hypothetico-deductivism and pure inductivism as opposite ends of the same spectrum, with cosmology and medical science representing their respective ends. And, in practice, both are probably oversimplified. My major point, I guess, is that there can be a massive difference between what a physicist and a biologist call “the scientific method.”