Science and the Art of Motorcycle Maintenance

In my last blog I was quite critical of studies that link climate change to the rise and fall of complex societies. I acknowledge that my critique was, in a sense, cheap, and not particularly productive. To make it productive I would have to come up with a viable alternative to the approaches I criticize. But such a viable alternative exists! It’s the good old Scientific Method.

The Scientific Method is an amazing tool; it even has a place in the art of motorcycle maintenance. At least so thinks Robert Pirsig, the author of this brilliant, but difficult book (I doubt that many people who bought it got much farther into it than the first 50 pages).

The logic of Scientific Method is fairly simple, although why we should do science in this way has been argued (and continues to be argued) by philosophers for at least past 300 years. It is also not simple at all in application. But the essence is straightforward enough.

The first step is to be very clear about the question we want answered. What is it that needs to be explained? (Asking the right questions is probably the hardest thing an aspiring scientist needs to learn.) Once we are clear on the question, we need to develop ‘hypotheses,’ which are simply alternative explanations. It is very important to have more than one explanation, because we can only judge how well a hypothesis does in relation to another possible explanation.

On the other hand, there is no need to try to cover all possible alternative explanations in one study. In fact, it is an impossible task – for any real-life problem there are effectively an infinite number of alternative hypotheses. Even dealing with relatively simple mechanical systems (which were designed by humans to be as tractable as possible), such as a motorcycle or a car, it is still very difficult to find all possible reasons why the damn thing is not working in any particular instance. I had a car like this (and most of my readers will probably be able to relate to this experience). After taking it to multiple mechanics, and each time paying more and more money, and never getting the problem solved, I finally got rid of it and bought a new car. Understanding how human societies function should be many orders of magnitude harder.

It was apparently the realization that the potential number of rival explanations is infinite that distracted Robert Pirsig from his studies when he was a biochemistry student, which had an unfortunate result, his expulsion from the university. Unlike Pirsig, I actually find this open-ended quality of science liberating – it would be a terrible thing for science to end by answering all questions (or, worse, become so routinized that we could turn it over to robots).

Anyway, once we have a well defined question and a set of possible answers, the rest is a matter of technique. The basic idea is to get enough data of sufficient quality that will decisively tell us which of the hypotheses fares much better than the alternatives. And just accomplishing this will be enormous progress (somethning that is rarely done in history, which is why in history explanations multiply, but are never cut down).

Of course we don’t know whether our best hypothesis, the one that explains/predicts data the best compared to others, is the ‘true’ one. But that’s OK. Unlike in matters of religion or ideology, in science we never can achieve certitude. There is always a chance that even a very good theory will turn out to be wrong and will be superseded by a better one. And that would be again progress, although we will still not have any certitude that the new and better theory is ‘true’.

By an interesting coincidence, today I went to a talk by David Archer, the author of The Long Thaw. David was talking about climate dynamics, although the main thrust of his research is not the effect of climate on history, but what it means for our future. During his talk, he cited one of the critics of the greenhouse effect, who said something like this, “You believe that the cause of global warming is the rise in CO2 emissions, because you are not smart enough to figure out the true reason” (I am quoting from memory, so if someone can point me to the original quote, I would be very grateful).

This is a great quote, and it demarcates perfectly the limits of science. Indeed, our best theories are only as good as our ability to dream up new explanations. Past experience shows that a scientific theory could do very well and explain a range of observations, and then turn out to be wrong (the paradigmatic example is the shift from the Newtonian to Einsteinian physics).

This blog ended up focusing almost entirely on method and philosophy of science.  In the next installment I will apply these rather abstract ideas to the concrete questions of how we can throw the full might of Scientific Method at the question of how climate dynamics influence the rise and fall of complex societies.