Skip to main content


Reductio ad absurdum

14 October 2014

I just can’t stop circling around the idea of reductionism. Perhaps because this philosophy is so central to all of modern science (and by ‘science’ I’m mostly referring to the life sciences, but I’m sure the principle holds true for other natural sciences as well). The reductionist idea goes back to the age of enlightenment, and so it is at the very core of what we understand to be the scientific method.  Virtually everything we do today in science is a reductionist approach to something greater, a model of a disease, trying to extrapolate universal principles of development from one species to another (sometimes as remote from one another as millimetre-long roundworms and humans), or studying the function of an organ by looking at its components. And the beauty is that these approaches work most of the time. Of course, the critical issue is to know the limits of the model system one is employing (and it is remarkable how many studies get published that employ the wrong models or the right models in the wrong way, so these limits are apparently not obvious even to some scientists). Choosing the right model is not trivial, and can be an art in and of itself (hence the title, no wait…). Because of its central role for modern life sciences, ‘the science of reduction’ deserves a closer look.

A  reduction is the process of boiling away excess liquid in order to concentrate the flavours in a sauce or a soup. Just checking if you are still paying attention. When I say science of reduction, I mean the complex task of anticipating which parts of the object under study are irrelevant to the question one is trying to address (and which can therefore be removed). That this is anything but trivial may be illustrated by the following example: suppose you have no clue about a new scourge of humanity called ‘cars’, killing several thousand people every year. In trying to understand why and how these ‘cars’ affect public health, you try to study how ‘cars’ work . You chose to use a reductionist approach for studying ‘cars’, because a ‘car’ as a whole seems awfully complicated, and quite frankly, one has to start somewhere. You would like to use a simpler model for ‘cars’, but unfortunately, there are no ‘car models’ available yet. So, let’s make a simpler model of a ‘car’. The tires look important, they seem to be relevant to moving the ‘car’ forward. So let’s keep them in the model. There are doors, and people seem to get into ‘cars’ using these, and sometimes these people die while using ‘cars’. On the other hand, some other people not in ‘cars’ die while coming into contact with a ‘car’. So, are the doors relevant?

I think you get my drift.