Saturday, November 6, 2010

The Darwinian "Theory" of Evolution

I have many pet-peeves. One of them is people claiming that scientific theories have been "proven." And I find that no more aggravating than with the "theory" of evolution. That is wrong on two counts:

  1. Scientific theories are NEVER proven.
  2. Evolution is NOT a scientific theory
The first one is quite simply a result of the nature of scientific investigation. Science moves forward by demonstrating old theories to have been wrong. No theory no matter how consensual can ever rest on its laurels confident that no-one can attack it. If you want proofs, go do math! But if you're doing science, you demonstrate. (On a good day)

As for the second, I would first like to provide a disclaimer. I believe in evolution. I love evolutionary game theory to explain and describe all sorts of phenomena. I think evolutionary biologists (or as my microbiologist friend says: "the computer scientists who wandered in my department") have the coolest job ever. I don't even believe in god or anything intelligent which would have or be creating or designing or guiding the world/universe/cows. But, evolution is not a scientific theory. Why do I say that?

Let us look at what scientific theories are. For something to be a scientific theory, it has to be falsifiable. People sometimes add on all sorts of other criteria but trust me, falsifiability can be found at the source of all other criteria. So what is falsifiability? To put it simply, it has to be conceivable to demonstrate the theory to be false. Is practice, what that means is there needs to exist an experiment (natural or artificial) which if it is performed and certain results come out, the theory can be pronounced false. For instance, I have the theory that: The sky is blue. It is possible for me to peek my head out, look up and see what color the sky is. If it is not blue, my theory was false. "The sky is blue" is therefore falsifiable. You will however notice that for that experiment to work, we need well defined terms. (Or operationalized variables as we sometimes say in academia) Let us consider that we define the word blue as being: "The color that the sky is." Well now, what my theory is really saying is: "The sky is the color that the sky is." That is tautological. There is no experimental outcome which would show that statement to be false. As an alternative consider the idea that we have not defined "blue" prior to the experiment. I could stick my head out the window, look and and say: "The sky is blue" no matter what I saw! In that case again, the theory would not be falsifiable.

Unfortunately, it is not enough to have an experiment. One must have a competing theory which predicts an alternative empirical result. The most basic competing theory is the negation of the theory: "The sky is not blue." However, competing theories can be quite complex themselves: "You live under a dome which never shows you the sky." If that theory was to be the competing theory, sticking your head out the door of your house would be insufficient. You must further demonstrate that the is no dome perhaps by constructing some sort of ladder or flying up in a rocket. This may seem silly but that is the very principle upon which science is built. A theory is always assaulted by competing theories which contend to explain more or better the universe around us. So for instance, Newton's theory of gravity was demonstrated to not be as accurate as Einstein's theory of gravity when a set of measurements were taken which better matched Einstein's theory than Newton's. So now that we are armed with knowledge and therefore power, let's look at evolution.

At its most basic, evolutionary "theory" simply says that random mutations occur in the population and those mutation which confer an advantage will result in the individuals passing on these mutations to its children thereby spreading those mutations. So how do we test that? Well, it is quite simple. One must simply look at a mutation that confers an advantage and see if that mutation spreads to the rest of the population after enough generations. (And do the experiment a thousand times in order to account for random and non-random factors) Easy? Well not so fast! How can you tell if a mutation confers an advantage? Evolutionary theory does not tell us anything about that. That is a crucial problem because after running my experiment and seeing that the trait does not appear to spread, I can safely claim that the trait was not actually advantageous. Perhaps it was disadvantageous. In other words, the variables cannot be correctly operationalized.

So now let's pit evolution vs it's greatest nemesis: intelligent design. Here, the resemblance in both stories is stunning. Intelligent design simply replaces natural selection and randomness with the hand of god. (or whatever other entity you choose to place) Of course, one cannot know what god wanted ex ante (latin for before the fact) so, whatever the outcome is can be safely described as being the result of the hand of god. Intelligent design and evolution fall in the same trap of not being falsifiable.